Literature Research Project



“Teaching tools for a new course in Robotics”

December 2015



The motivation behind this project was to generate an impetus for a new Robotics course for the Department of Mechanical Engineering in Imperial College London. The author had intrinsic motivation regarding this project because of his interaction with robots and unmanned vehicles in the Singapore Armed Forces.

The problem statement was strategically deconstructed and the case for robots and robotics first developed, as well as explanations for their relevance to mechanical engineering and the world. Consequentially, the stimulus for teaching robotics arose from the demand by the robot industry for engineers with the requisite expertise. The current situation was briefly researched, both at tertiary and pre-tertiary levels (since the course would eventually be taken up by present-day pre-tertiary students).

The notion of robotics as a module was hence developed and core components identified and the rationale behind each component established. The tools for teaching robotics were then explored in depth – starting from the fundamentals: learning outcomes, going down to the learning experience in terms of venue, hardware and software resources, learning methodologies and finally assessment methods and their relationship with learning.

The conclusion underscores that there leaves much to be explored, and highlights the amount of resources and dedication required for a course to be purpose-built for the department, but also stresses the positive effect on industry, and subsequently, society.



1.1       Objectives

The objectives of this LRP were to explore the case for a Robotics course in Imperial College London’s Department of Mechanical Engineering, and conduct research to develop course structure, and tools and methods for teaching robotics. The LRP should provide insight and material useful for consideration when kick-starting the Robotics course.

The nature of this LRP is different and not wholly about engineering concepts, hence the diverse nature and heavy reliance on internet sources. Some portions of text will seem to lack citations because these are either from the author’s own knowledge or are claims that can be generally accepted to further develop the argument, or present logical inferences (or lead to them) within context.


1.2       The Case for Robots

Exponential advancements of technology in the last five decades, as predicted by Moore, can be attributed to a decrease in costs, size, speed and reliability.[1] A reduction of time, effort and cost for production accompany these developments.[2] Mankind has strived to find substitutes that can mimic human behaviour.[3]

The poignant driver of progress in robotics is minimising the need of humans to perform repetitive and menial tasks which can otherwise be automated, hence freeing humans to channel their energy towards creative and critical thought. A recent study by Oxford University addresses the pressing question, “Will a robot take your job?”. Nine key skills are identified and quantified that determine the answer to this question. The primary notion is the difficulty of automation. For example – social perceptiveness, negotiating, and the fine arts are hard to manipulate, but as technology improves, dexterity and complexity are easily tackled by the robots available today.[4]

Another driver that has risen to significance particularly in the last few decades is precision engineering. The ability of robots to sense and actuate with small error tolerances, coupled with the facility to execute dangerous tasks in hazardous conditions with minimal mistakes, explains their enhanced role in manufacturing currently.[5] In quality control, there may be tasks whereby humans may outperform robots in either speed or quality, but robots perform superior in terms of consistency,[6] which is a primary consideration for many manufacturing firms.

Economically, a robot will pay for itself eventually: the increased productivity of introducing a robot outweighs its cost.[7] This cost-benefit analysis is a function of time, complexity and value.


1.3       Present-day Considerations

History has conceptualised robots with a number of ideas focusing on a central motif. The early term for ‘robot’ was ‘automaton’. Modern robots were first adopted by factories to enable production independent from human input or assistance.[8]

Today, the forefront of robotics research lies in materials handling and manufacturing, cooperative and swarm robotics, human-robot interactions, robot ethics and design.[9] Robots of today can teach each other new tricks, self-replicate, self-repair, self-upgrade, self-organise and collaborate, be independently modular, and survive indefinitely despite being potentially self-sacrificial.[10]

Robots are quickly being embedded into modern warfare and military tactics. Development of driverless cars is well underway. Even tasks which, based on the study mentioned in §1.2, are unlikely to be automated, have been slowly been taken on by robots. Drones are taking over service at some restaurants, and Humber River Hospital is ‘fully digital’.


1.4 Defining ‘robots’ and ‘robotics’

Robots are machines that are designed to follow a computer programme to complete a specific task list. There are four levels of autonomy in robots:

  1. Remote controlled: the robot is programmed to respond to specific inputs from the end-user to actuate some mechanical response, with little or no sensors. There is no defined task-list aside from the immediate demand of the end-user.
  2. Pre-programmed: the robot completes the task list as specified based on parameters given prior and purely proprioceptive sensors, without further input from the end-user.
  3. Autonomous: the robot makes decisions on its own based on the computer programme and inputs from the environment that it senses and analyses itself
  4. Artificial Intelligence (AI): the autonomous robot that also learns from its environment and can modify itself to adapt to it and complete a task list in the most efficient manner

Robotics is the field of engineering concerned with robots.



2.1       The Case for Robotics

Present-day engineering cannot be taught without some reference to robotics. Engineering and robotics move parallel in development because both strive to achieve the same ends of easing the workload of man. Robots pervade mankind’s existence; to control, fully functionalise and develop robots to maximum utility requires a keen knowledge of robotics.

Robots have replaced the mechanic but not the engineer. Engineering is still in high demand because of the skills and knowledge required for design and practicality, as well as creativity which robots lack due to their encoded rigidity. The modern engineer will turn to robots for their competency and consistency, and use robots and their knowledge of robotics extensively to solve problems. Interactions between engineers and robots will grow as the world becomes reliant on robots to serve heavy-duty needs. Close contact necessitates a comprehension of the technology one finds oneself dependent upon. Without a doubt, an engineer would minimally need to have an appreciation for robotics.

As the world moves towards greater automation, so too should engineering. As the industry develops, the amount of expertise required will increase proportionately. Engineers need to synergise with robots in order to increase productivity, as well as remain relevant to shifts in the technological paradigm.


2.2       Teaching Robotics

Epistemically, in the United Kingdom, the first interaction schoolchildren have with Science, Technology, Engineering and Mathematics (STEM) is through Robotics.[11] Nurturing talent early in a field of pertinence from the outset is a solution to the engineering shortage facing the UK.[12] Economically, the market’s appetite for engineers should drive the teaching of robotics from a young age. This, together with the thrust towards increased automation, creates sufficient incentive for robotics to be included in pre-tertiary curricula.

It thus follows that at the tertiary level, the requisite learning and knowledge in the sphere of robotics drives a demand of its instruction and knowledge management, evidenced in the preceding section. Robotics is an interdisciplinary field, leaning most towards mechanical engineering, incorporating mechanical design in the robot’s external structure, mechatronics in its internal hardware, and programming in its software makeup. Hence, an engineering education in robotics would require merging some of these elements. The lines between different areas of engineering have been greyed, and robotics is a causal factor. Cross-faculty understanding is crucial for engineering and robotics to progress.

Mechanical engineering courses have elements of engineering design, mechanics, mechatronics and control systems, as well as computing/programming.[13] This engenders an impetus for a mechanical engineering course to offer a module in robotics. In all of the top ten universities (four of which are STEM universities), degree programmes or courses/modules (used interchangeably in this report) in robotics are offered, and if classified under a department, these modules are taught by the respective mechanical engineering departments.[14]

Beyond the realm of engineering, robots also compel anthropologists, sociologists and philosophers to explore ethical considerations, synergistic relationships between robots and humans, and the shifting dynamic between humans and robots. These ideas have been explored in science fiction, but leave room for rigorous expansion. Siciliano et al. expounds on the history and ethics of robotics in their similarly-titled book; this can be a starting point for engineers and non-engineers alike in their understanding of robotics.



3.1       Core Components

Robotics as an engineering module requires a right balance of the components that make up robotics. Robotics straddles between its hardware and software, and one of the foci should be integrating these two symbiotic parts. Since humans are masters of technology, another focus should be the harnessing and controlling the technology to employ them for use.

There is no literature on the quantification of proportions of these core components, but this section aims to explore the components individually and provide insight with regards to the two aforementioned foci, as to how to structure a module in robotics.

As a precursor to the practical aspects of robotics, there are certain concepts which can be introduced, enhancing the understanding of robotics. Such notions are relevant but non-technical and include (non-exhaustively):

  1. The stimulus and motivations behind a module in robotics
  2. The history of robots
  3. The ethical implications in robotics
  4. Intelligent design and artificial intelligence


3.2       Mechanics   

Kinematics, mechanics and dynamics all refer to motion and is of interest because a key feature of robots is locomotion; at all levels of autonomy, the ability for a robot to propel or drive itself or one or more of its particular sub-mechanisms in a specified vector characterises robotic movement.

The Massachusetts Institute of Technology (MIT) offers a 2.12 Course titled ‘Introduction to Robotics’, which falls under their mechanical engineering department. The course ‘provides and overview of robot mechanisms, dynamics and intelligent controls’. Students work on a term group project, and have weekly laboratory sessions.

Students in MIT are also assessed through an exam with a heavy emphasis on spatial dynamics and multi-rigid-body dynamics. Key words such as ‘forces’, ‘torque’, ‘inertia’, and ‘Jacobian matrix’ appear in problems that revolve around robots and robot designs.[15]

The fact that particular attention is paid to dynamics in the exam from arguably a leading university in this field, demonstrates the importance of understanding mechanics principles before an attempt at mechanical design is made. It follows that the coursework component focuses on control design, embedded software, task modelling and the human-machine interface. The case for robotics as an examinable subject will be discussed in §4.3.

The California Institute of Technology (Caltech) Introduction to Robotics course begins with and spends about half of the whole course studying robot kinematics and odometry.[16]


3.3       Mechatronics and Control

Basic mechatronics finds its place in robotics because it serves as the bridge between electronics and the actuated mechanical motion, through servomotors or their equivalents. The electric component runs the power supply, the power transmission to the actuator and the power supply to the electronics.

Control theory finds itself steeped in robotics as well. When the machine tool industry dominated robotics, each of the robot’s joints was controlled independently as a linear single-input/single-output system, and exteroceptive sensors were non-existent. Now, after much headway in control technology, especially automatic control and the reduced costs and increased speeds involved in computing, control theory is synonymous with robotics.[17]

Courses in control would succeed in representing the condition of robotics and the overlap of mechanical and electrical systems, even if there is no explicit reference to robotics in the course material. ETH offers a ‘Master of Science in Robotics, Systems and Control’, which illustrates the speciality of such a field,[18] and also the departments jointly offering this course are of the fields mentioned in §2.2.

In today’s cars, up to the last decade, automatic parking is a fast-growing notion that is also quickly gaining commercial conception and epitomises the application of control theory in autonomous systems (in this case, a car that parks without guidance from the driver).


3.4       Programming

Programming a robot is the means by which engineers communicate instructions to the robots,[19] and the same way robots communicate with each other in collective robotics. This is a fundamental skill in an engineer’s ability to control and employ a robot to complete a specific task list. Robots respond unilaterally to programming language and enables a robot to respond in the most precise manner it is commanded to, rarely making mistakes.

The way many university courses teach programming is to introduce programming concepts, either by explaining how the computers interpret these concepts or providing analogies to explicate the meaning of the code. Students are then required to tackle challenges which force them to manipulate these concepts and apply the logic in the language to find solutions to these problems.

Coding a robot presents a steep learning curve, but is the baseline for maximising the potential of a robotic system.[20] Robotics cannot be taught without some form of programming, no matter how basic or mature it is. It does require knowledge of computers and an understanding of how adjusting various parameters delivered to the robot will result in different responses.

Programming ties in significantly with control, because many of the parameters in control theory can be manipulated through software and programming, hence while control theory forms a basis for uses in robotics, the practical application lies in experimentation and laboratory-based learning.[21]

The future of programming lies in the ability to manage cooperative and swarm robotics – robots that work together to achieve the same singular purpose.[22] For an introductory level course, this can only be appreciated without in-depth understanding of the logic fundamental to the programming of such groups.


3.5       Engineering Design

There are modules with intense focus on engineering design in many curricula across universities. Vital aspects of such courses explore the design process, Computer-Aided Design, design against mechanical failure, and design for manufacture. More importantly, design is considered a core component of robotics because of the creativity in finding a way to automate processes or solve problems in robotics.[23]

Design in robotics encompasses the incorporation of software into hardware, as well as constructing mechanical systems and mechanisms that can be controlled electronically to achieve the intended outcomes for the robot. There is much theory behind design in robotics; often this is not elucidated directly, but instead through concepts in other core components. For example, an examination question in the MIT robotics course asks about a ‘two degree-of-freedom planar robot with a parallelogram mechanism,’[24] a common design adopted by many robotic arms in a factory-line.


Image 1: Parallelogram mechanism for robotic arm[25]


Design should be learnt by exploring creative solutions to problems, such as overcoming high weight-to-load ratios, or maintaining certain spatial orientations while being constrained in certain axes.



4.1       Learning Outcomes

A majority of the introductory level courses to robotics end with a challenge or specified task list, which constitutes a significant proportion of the course grades. This serves as a fitting mode of assessment, as well as accumulation of all the required skills accompanying an engineer who has elected to take a course in robotics. With this end-state in mind, one can back-track to develop a course that ensures a student is able to pool his knowledge of robotics and tap on each of the core components to build a robot. This allows one to properly allocate proportions of the syllabus to meet the requirements of each constituent component, before the course converges on the project.

In order to introduce an entry-level robotics course, the course structure must be varied and balanced in all aspects and thereby develop learning outcomes in all realms to varying degrees. Bloom’s taxonomy is a good model to adhere to when developing learning outcomes,[26] and learning outcomes can be set out to ensure sufficient depth (up the triangle) is achieved.


Image 2: Bloom’s Taxonomy[26]


A successful introductory course would develop all core components to minimally a degree of understanding and light application, and students should develop an appreciation for the non-technical aspects of robotics as mentioned in §3.1. Also, the course should prepare engineering students for more advanced courses in robotics, and be able to put the student in good stead to deal with robotics as an industry professional.


4.2       The Classroom/Laboratory Setting

The Setting

The nature of classrooms (lecture theatres) as confined spaces with theoretical principles needed to be communicated within a limited time, and the nature of an engineering course as a downstream discipline of the hard sciences with laws and rules that govern physics, leaves little space for formulation of ideas and discussion. The classroom is effective in this regard.[27] However, not going beyond the classroom risks hampering creativity and limiting the development of problem-solving skill sets that are crucial to an engineer.

This is where the input of a laboratory is essential in the learning and thought processes of the engineering student. The laboratory has long been considered to have a ‘distinctive role in science education.’[28] Learning in a laboratory has been claimed to develop different groups of objectives, on top of reinforcing concepts, as illustrated in Image 3 below. The laboratory gives precedence to first-hand experience, and ‘the activity of the student, sensorimotor nature of the experience, and the individualisation of laboratory instruction’ should contribute positively to learning.[29]


Image 3: Different objectives of laboratory learning


In the safety of a laboratory under controlled conditions, students can explore and tweak the parameters of the robot’s code to prompt different outcomes. This hands-on approach promotes PBL and incorporates many characteristics aligned with the spirit of engineering, such as innovation, teamwork and tenacity, that would be difficult to cultivate in a classroom setting.

Aside from head knowledge, other skills can be imparted in the laboratory. Laboratory etiquette is highly emphasised on in many universities and is relevant even in an electronics laboratory at university level. Acquiring the attitudes as in Image 3 from good laboratory practice will have ripple effects in giving students ownership of their learning.[30]

The classroom and laboratory complement each other, where information should first be presented before its applications explored in the laboratory. Giving lectures in the laboratory immediately prior to the start of laboratory activities could prove in this light to be most effective in knowledge retention and near-instantaneous application. Lectures in the safety of a laboratory can incorporate demonstrations which may not be possible in the classroom setting.

By timetabling laboratory sessions straightaway after lectures, the lecture’s effectiveness is elevated. This is supported by Kember’s notion that knowledge is constructed by students and that the lecturer is a facilitator of learning rather than a presenter of information.[31] The notions of discussion, practice doing, and teaching others also tie in well with the objectives of a laboratory setting in Image 3.

Laboratory sessions often come accompanied with a set of laboratory handouts. Some require the students to fill in blanks and complete tasks to be checked by tutors, while others serve as precursor reading to the laboratory session. The laboratory handout is a vital tool in guiding and promoting learning for the student. The handout should not try to re-teach theory taught in the classroom, but be supplementary, the same way the two settings are symbiotic in a student’s learning. One has to be wary that the ‘primary value of the lab is in the act of investigating, not simply in the result of the investigation’,[32] and hence, the handouts should focus on processes and working towards an end-result, thus encouraging learning and not rash fulfilment of tasks.


This is usually followed up by the student being required to furnish a laboratory report to report their findings but also demonstrate understanding and application,[33] and moreover summarise learning. The nature of such laboratory reports is therefore archival rather than persuasive.[34] The frequency and required length of lab reports should be considered together with other workload from the course, to ensure learning is effective and unrepeated, and to guarantee a firm record of learning. This is considered again in §4.5.

A laboratory seems infinite in its ability to render possibilities, but an introductory course in theoretical and practical aspects of robotics can only be taken so far. The limitation of time is a function of the constraints of the provision of a laboratory, and an introduction to robotics would need to channel resources to a specified breadth of knowledge and less so depth. The idea of knowledge management is central in the formulation of a course structure as mentioned in §4.1.

This implies that the course cannot attempt to delve too deep into a certain area, or even expect extrapolation of such knowledge to be applied in the final challenge. Take locomotion as an example – planar motion should be easy to manipulate, but extending this idea to the third dimension is exponentially harder to deal with. Even in the core component of mechanics alone, a third dimension causes a doubling of the number of degrees of freedom. This is incommensurately difficult to programme and control for a preliminary robotics course. One can theoretically convey the complexity of robotic design in three-dimensional space, but this should only be left at the ‘appreciation’ degree of knowledge management – not even on Bloom’s taxonomy!


Group Size

There is widely varying literature on the size of groups for ideal learning, but there is ‘no hard-and-fast rule to determine the optimal number’ to have on a team.[35] One can consider the study published in Scientometrics in 1980, which found that publication rates had a positive correlation with research group size.[36] However, knowledge generation and learning – which is more a form of knowledge management – are different. Also a large group cannot be expected to have equal rates of learning, and hence a larger group does not correspond to more learning.

Edward Hall, a psychoanalyst and anthropologist, concluded that the perfect group size is 8-12 people. This is natural due to our evolution as primates living in small groups, and ‘participation and commitment fall off in larger groups’.[37] Also, logically speaking, one large group working on a single bench will not ensure a fair distribution of knowledge management. The numbers vary widely, and different universities will present a number closer to 4-6 for projects and this can be extended to the laboratory setting, especially if projects are incorporated into laboratory learning, so all member can contribute. To overcome this issue of managing learning in courses of large group sizes, a course leader can always require completion of individual handouts or laboratory reports.

However, it must be argued that earlier on it was suggested that there are merits to the individualised learning that occurs in laboratories. This is in conflict with the ideas of teamwork and communication which should be developed in a lab, although admittedly gives more time for self-practice, which in turn increases knowledge rates. A compromise for this line of contention would be pair-work or at most groups of three. In fact, Nobel prizes in Physics and Medicine have been shared between 2 or 3 laureates more often than have been awarded to individuals,[38] which show the benefit of working in groups in the realm of scientific advances.

However, group size is recognised to be a function of resource quantity as well, and groups may need to be bigger to balance and divide resources between all students properly. This should never be at the expense of a student’s access to learning and sufficient resources should be pumped into a course to ensure this.


Staff-Student Ratio

It is important, considering all of the above, to also explore the staff-student ratio – laboratories and classrooms are not only limited by physical capacity but also by the ability of academics and teaching assistants to guide and track students enrolled in such courses, as determined by researchers from the University of Oklahoma.[39] A particular number of students pursuing a robotics course requires a corresponding number of staff able to teach this course, which dictates that a pre-cursor to initiating a basic robotics course is having enough academics at the graduate level whose focus is related to robotics such that they can be effective in instructing the course.

There is no literature on an ‘optimum’ or ‘ideal’ staff-student ratio (SSR); even for generic university-wide or department-wide SSR numbers, no specific quota is set. It is probable that for laboratory sessions on robotics courses this will vary widely, between different universities/departments but also between different year groups in the same department. A more suitable indicator would be the time a student has to wait before he is attended to by a tutor. The course leader can set a benchmark and adjust the SSR if students are waiting too long, or if tutors are standing around too long.

In essence, laboratory time has to be maximised to allow students to explore on their own, while ensuring the course structure and learning objectives are well-tailored to cover enough scope to a given degree of knowledge, with sufficient resources so every student has space to explore and develop.


4.3       Technological Tools

Virtual Laboratories

In the discussion in §4.2, the notion of the laboratory has been in the physical sense, however, there has been research done in the recent past into virtual laboratories for science and engineering education, which could be assimilated into a new robotics course. Virtual laboratories can ‘simplify learning by highlighting salient information removing confusing details’ and require less set-up time, providing results instantaneously. Logs of student work can also be used by teachers to ascertain which students require more help and hence ‘refine their instruction’.[40]

The most glaring benefit of having a virtual laboratory for an electronics course is that there is no issue with hardware at all, but conversely, there is no tactile contact with the robot. Studies suggest that this is not a requirement for the development of conceptual knowledge except students who do not have previous relevant physical experience.[40] For an entry-level course, virtual laboratories may hence be less suitable.

This said, virtual laboratories could still be advantageous for programming-heavy courses. For courses with a focus on mechanics and control, the virtual lab will not be effective. Take, for example, a simple task of a robot picking up a ball. A computer programme would be very complex to create and manipulate to test the mechanics of a virtual robot in its ability to complete the set task, and a physical robot picking up a physical ball will realise the intent of the exercise in a superior fashion. A balance can be struck to capitalise on the features of each approach.[40] For example, the locomotion of a robot can be tested virtually to ensure that there is no issue with the software before physical testing to ensure the physical robot can achieve its objective in reality.


Audiovisual Learning and Demonstrations

Learning also can be enhanced with audiovisual learning and demonstrations.[41] A programming heavy course can show videos with analogies of what a particular function or class of commands does for a given programming language. A mechanics-heavy course can show videos of expected outcomes of the robots – even better if these videos are taken from best practices of a previous year’s conduct of the same course because it proves to the students that it is not a far-fetched expectation, and challenges students to find even better methods of doing completing a set task.

This serves as a stepping stone for students’ further development on their own.[42] The issue at hand when considering imitation of robotics concepts is the ability to understand the code and concepts and manipulate them to suit other engineering problems. This is an antecedent to the concept of grades and learning explored in §4.6.

The provision of videos also helps to define a task more comprehensively, and closes it to alternative interpretation. However, this may have a stifling effect on creativity in the spirit of Problem-Based Learning (PBL). In some cases, this may not be an issue as long as the task is explored to sufficient and equivalent complexity, but in many other cases in robotics and engineering, the task will be very specific and explicit and leaves no room for interpretation.


Software and Hardware

Since one of the core components of robotics is programming, having proper hardware and software is important. Without saying, a student should have access to a computer with a powerful enough processor to support the software development, and also should have access to all the tools to manipulate the robotics hardware accorded to them.

For an entry-level course in robotics, it is necessary to start from other-than-bottom as argued in §4.2, and hence, the programming language used should be sufficiently high-level so as to ensure the focus remains on programming the robot to complete the task, unless the robotics course is desired to be programming-centric. It is accepted that robot software is highly proprietary in nature and comes accompanied with robot hardware. This lack of standardization poses a problem to learning, but all the languages are quite similar and have syntax based on a lower-level programming language.[43]

As an example, the software accompanying LEGO Mindstorms EV3 is relatively high-level. This is necessary to keep things simple since Mindstorms is used extensively in primary and secondary school robotics. The language is simple – different icons represent the actuation of different parts of the robot, and these icons are arranged in a sequence of commands for the robot to execute. This nearly removes the programming aspect of a course in robotics. Nonetheless, there are tools available online developed to allow the LEGO robot to be manipulated using text-based programming (that is, taking the programming down to a lower level), which will be more suitable for a university introductory robotics course.[44]


There is no exhaustive or definitive software to adopt to teach robotics, but it is probably advisable to use a language which can be easily understood, learnt, and applied and be low level enough to transcend robotics to be manipulated to other engineering applications. This will also provide the course with a tangible link to engineering. There are many papers based on academics in computer science who develop programming environments for teaching robotics, and the keywords for such papers are ‘education’, ‘robotics’, ‘control’, ‘programming languages’, and ‘computer science’.

These environments described by the papers use a variety of languages and using one’s desired language as an additional keyword should display a paper explaining the environment created. It is much easier and consumes much less resources than developing proprietary software for a robotics course.

Regarding hardware, there are similarly many options available on the market for teaching robotics. The programming environment must be appropriately chosen to suit the hardware and it is easier to check for compatibility this way than vice-versa. This is because software can be manipulated to actuate motion but if the hardware is not able to perform this function due to a hardware limitation then the robotic system will be moot.

LEGO Mindstorms are highly popular in teaching robotics especially from basics, because they encourage creativity, and in the spirit of an introductory course, remove the need for starting-from-scratch regarding materials and hardware (and following this line of argument are also easy to build). They also are of high quality and are available and accessible worldwide, have a good support system and are well-documented, and easy to assemble, disassemble and reuse to build other things.[45] Although LEGO Mindstorms seem to be the predominant choice – and for good reasons too – there are alternatives on the market.

Image 4 shows different alternatives to LEGO Mindstorms. Some, like Makeblock,[46] are dynamic enough to be explored in a sufficiently broad and deep manner. These systems also have an obvious emphasis on mechanical motion in their systems design and hence will be suitable for a robotics course based in a mechanical engineering department. There are other simpler systems on the market such as Cubelets,[47] but these kits are not as comprehensive for a university-level course.

Image 4: Ideal (left, MakeBlock) and non-ideal (right, Cubelets) alternatives to LEGO Mindstorms

There are many advancements in the robotics industry[48] that cannot be explored in a robotics course because of the constraints of time, money, and expertise amidst other resources. These include swarm robotics which are generally proprietary in nature and can be explored at other levels other than undergraduate. Arguably this will make any university attempting to teach such modules a pioneer, and will blaze the trail for others to follow. Naturally this has positive effects for a robotics course at the basic level as the course can and will be driven forwards by advancements in this area done by in-house faculty.

Other areas include AI: this is a field which is very programming-heavy and can be introduced to students but is too high-level to be integrated in an entry-level course, which can instead include Asimov’s three laws of robotics (to the ‘remember’ level of Bloom’s taxonomy). Again if research is pioneered by postgraduates it will serve to help develop the robotics course further in this field. AI can be taught instead by considering the next-lower level of autonomy as described in §1.3: autonomous robots which use exteroceptive sensors to take in information on its surroundings, and analyse the situation before making a decision based on pre-input logic. This is more accessible to the entry-level student.

Ultimately, the tools employed will vary depending on the preference of the course leader in accordance with the pre-set learning outcomes, and universities can take the initiative to develop such programmes with senior students and researchers specialising in this area.


4.4       Coursework and Experimentation

The University of Oklahoma study elaborated on the instructional strategy of choice, where PBL was the chosen approach to supplement the ‘traditional lecture’. PBL is an educational psychology scheme where students are presented with a problem to which they have to find a viable solution.[49]



Image 5: The PBL process


The students have both the physical space to explore and innovate, as well as considerably less mental pressure and time limitations than under examination conditions. There is room for error while finding the optimum solution to any problems pertaining to hardware or software, with guidance and without assessment.

PBL give students an objective to work towards and in the process cement robotics concepts, and build on abilities and knowledge, and learn ways to solve problems.[50] Especially in the realm of programming, there are multiple ways to solve a problem; one method may be significantly more efficient and instructors step in to clue students towards this method.

While this is in tandem with the idea of PBL as an ‘open-inquiry approach’ – where learners have time to ‘struggle and define the problem, explore related issues (during and/or after sessions), and grapple with problem resolution’[51] – this is difficult to achieve under timetable constraints.

The limitation of robotics coursework is the difficulty in starting any form of robotics from scratch – a base needs to be set-up for learners to build upon. The University of California, Davis (UC Davis) study elucidates the idea that the process is time-inefficient, but more dangerously could lead to erroneous conclusions and provides little guarantee that the material learnt through PBL can be applied. The paper compares PBL with Case-Based Learning, and argues that the ‘presence and expertise of the faculty is wasted if not harnessed in more than a passive manner’,[51] which is in line with the assertions in §4.2.

Hence for effective use of time in the lab, the hardware should be provided at a certain level of construction and a skeleton code formed so that students need not make large leaps in learning to support their understanding of robotics. Students should be cognizant of this fact and merely appreciate that there is much prior planning and engineering from where they begin.

The laboratory should not be a highly secure place for a course such as electronics involving low-level risks and threats and no hazardous materials – and this gives the laboratory an extended atmosphere of open learning while never compromising on safety. Universities often leave labs open during workdays and restrict access when courses are in session, which allow students to work on both course-related and extracurricular projects in the laboratory. This encourages learning and develops many of the points in Image 3. It also underscores the nature of engineering as a professional practice.[32]


4.5       Examinations, Grades and Learning

The effectiveness of an examination for robotics is severely limited. It is difficult to examine an understanding of programming unless an exam is computer-based, and this undermines the spirit of computing by forcing it in a very rigid structure and setting time constraints. There is little literature on examinations and assessment for robotics courses, which perhaps is suggestive of the lack of examinations in robotics. Hence, this section was written by empirical evidence of courses currently offered in the top universities worldwide, in particular STEM universities.

Courses in Robotics offered by various universities hence have a large proportion of course grading allocated to coursework, which demonstrates the emphasis of practical work and hands-on experience that robotics courses entail. While a robot completing the task list may be very telling quantitatively of the students’ grasp of robotics, it is more important to get right the robot’s implementation and execution. As mentioned in §4.3, there should be a focus on the ability to understand concepts in the course and manipulate them to suit other engineering problems. Due reward should be given to completing a task in the most efficient way, no matter to what degree one uses hardware and software to achieve this.

The robot need not be perfect, but the intent behind the execution of the given tasks should be firm – students should strive to ensure their robot is working towards this. After all, the converse is true: what seems like a creative idea to tackling a task may not be tested exhaustively and may fail in another way when the robot is subject to minor modifications as above. If equal weightage is given to creativity and evidence of attempt to adhere to task list, this leaves room for students to expound and develop, and they should be commensurately rewarded as long as no dead-end is hit.

Next, although the aim of the robotics lab is to actuate motion in the robot, the neatness of the software and hardware (the code and the robot’s structure) also has a degree of importance. This should also be assessed and the course leader should determine to what degree each of these factors are weighted.



5.1       Starting a course

Starting a course requires, besides the hardware and software mentioned in §4.3, prerequisite knowledge and expertise in robotics. Also, to differentiate this robotics course from other courses, creativity is required to find a niche area that the university can specialise in. It would be ideal to have a good base of academics (postgraduate students and faculty) who have a predisposed interest to kick-start the course. Good tutelage will ensure that a virtuous cycle of generation of interest and research in the area propels the course forward.

Also, the idealism needs to be tempered with pragmatism: there are entire postgraduate degree programmes dedicated to robotics, so one cannot expect much from an introductory level course especially in its infancy. Both faculty and students have to be aware of this and prepared for it, and be ready to offer suggestions to improve the course and help.


5.2       Further Developments & Evaluation

The same way mechanical engineering is applicable to robotics, so robotics remains extremely relevant to mechanical engineering and will take us to the space age (if we are not already there) and beyond. Advances in manufacturing and the medical fields will also benefit mankind. Beyond what has been explored in this paper, many more stones are yet unturned. One upcoming field is the idea of cross-training, so humans understand robots better and vice-versa.[52] This cannot be explored in an introductory level course, but is a possible niche area. MIT is currently the go-to on these matters.

Writing this paper definitely would cause inadvertent personal biases of the author to be included. This is a non-issue when exploring ideas, but the author could be predisposed to particular technology or methods, or in his study of mechanical engineering could be led to view robotics in a particular perspective.

In conclusion, an introductory course for robotics looks challenging to fillip, but with adequate backing in resources and expertise, and with a clear vision of the aims of the course, a mechanical engineering department will stand to benefit from a new robotics course, which will develop a generation of engineers ready to spearhead and develop the robotics industry.



  1. Takahashi, Dean. “Forty years of Moore’s Law.” Seattle Times, April 18, 2005. Accessed 3 November 2015.
  2. Butte Patil, Partik Rajendra. “Technological Advancements and Its Impact on Humanity.” B. Tech thesis, Mahatma Phule Agricultural University.
  3. Siciliano, Bruno, Lorenzo Sciavicco, Luigi Villani, and Giuseppe Oriolo. Robotics: Modelling, Planning on Control. London: Springer-Verlag, 2010.
  4. Frey, Carl B., and Michael A. Osborne. “The Future of Employment: How susceptible are jobs to computerisation?” University of Oxford, 2015.
  5. “Benefits of Robots.” Accessed 3 November 2015.
  6. ST Robotics. “Why use a robot?” 1997. Accessed 3 November 2015.
  7. W.. “Robots and social justice.” 8 March 2011. Accessed 3 November 2015.
  8. “A Brief History of Robotics.” Accessed 16 December 2015.
  9. MacIntosh, Anna. “Sheffield drive forward robotics research as part of new UK network.” 24 June 2015. Accessed 3 November 2015.
  10. Dvorsky, George. “12 Reasons Robots Will Always Have An Advantage Over Humans.” 16 December 2014. Accessed 3 November 2015.
  11. Pursuit of Unorthodox Ideas. “The Role of Robots in Engineering Education.” 26 October 2013. Accessed 3 November 2015.
  12. Onita, Laura. “Shortage of engineering skills to cost UK £27bn a year.” 12 January 2015. Accessed 3 November 2015.
  13. Columbia Engineering. “What is Mechanical Engineering?” Accessed 16 December 2015.
  14. QS Quacqracelli Symonds. “QS World University Rankings® 2015/16.” Accessed 3 November 2015. search=\
  15. Massachusetts Institute of Technology Department of Mechanical Engineering. “2.12 Introduction to Robotics: Exercise problems and Solutions for the End-of-Term Exam.” Accessed 3 November 2015.
  16. Ma, Jeremy. “Introduction to Robotics.” Accessed 3 November 2015.
  17. Spong, Mark W., and Masayuki Fujita. “Control in Robotics.” Accessed 3 November 2015.
  18. ETHzürich. “The Program.” Master of Science in Robotics, Systems and Control. Accessed 3 November 2015.
  19. Marrer, Gary. “Fundamentals of Programming: With Object Oriented Programming.” 30 November 2009. Accessed 16 December 2015.
  20. Benson, Coleman. “How to Make a Robot – Lesson 10: Programming Your Robot.” 29 June 2012. Accessed 16 December 2015.
  21. Luenberger, David. “Mathematical Programming and Control Theory: Trends of Interplay.” Accessed 16 December 2015.
  22. Barbato, Massimo. “Swarm Robotics: The Future of Collective Intelligence.” Accessed 16 December 2015.
  23. vexEDR “1.3: What is the Engineering Design Process?” Accessed 16 December 2015.
  24. Massachusetts Institute of Technology Department of Mechanical Engineering. “2.12 Introduction to Robotics” Accessed 3 November 2015.
  25. The Steampunk Forum. “Designing drive mechanism for large writing machine.” Accessed 3 November 2015.
  26. Overbaugh, Richard and Lynn Schultz. “Bloom’s Taxonomy.” Old Dominion University. Accessed 16 December 2015.
  27. CIRTL Network. “Lecturing: Advantages and Disadvantages of the Traditional Lecture Method.” Accessed 16 December 2015.
  28. Hofstein, Avi. “The Role of the Laboratory in Science Teaching: Neglected Aspects of Research.” Review of Educational Research 52, 2 (1982): 201-217. Accessed 3 November 2015. doi: 3102/00346543052002201
  29. Blosser, Patricia E.. “The Role of the Laboratory in Science Teaching.” NARST Research Matters – to the Science Teacher 9001 (1990). Accessed 19 November 2015.
  30. Fletcher, Adam. “The Architecture of Ownership.” Association for Supervision and Curriculum Development 66, 3 (2008). Accessed 19 November 2015.
  31. Kember, David. “A reconceptualization of the research into university academics’ conceptions of teaching.” Learning and Instruction 7, 3 (1997): 255-275. Accessed 19 November 2015. doi:10.1016/S0959-4752(96)00028-X
  32. the Physics Classroom. “Purpose Driven Labs.” Accessed 19 November 2015.
  33. University of Toronoto. “The Lab Report.” Accessed 19 November 2015.
  34. Dolphin, Warren D.. “Writing Lab Reports and Scientific Papers.” Accessed 19 November 2015.
  35. Wharton University of Pennsylvania. “Is Your Team Too Big? Too Small? What’s the Right Number?” 14 June 2006. Accessed 19 November 2015.
  1. Cohen, J. E.. “Publication Rate as a Function of Laboratory Size in a Biomedical Research Institution.” Scientometrics 2 no.1 (1980): 35-52. Accessed 19 November 2015.
  2. “Edward Hall: The perfect group size = 8 – 12 – Signal vs. Noise (by 37signals).” Accessed 19 November 2015. http://robertogreco.soup. io/post/58843180/Edward-Hall-The-perfect-group-size-8
  3. org. “Nobel Prize Facts.” Accessed 19 November 2015.
  4. Annesley, Timothy J., and David P. Miller. “Review of a Robotics Lab for Engineering Undergraduates.” Paper presented at the 2008 meeting of the American Society for Engineering Education, 2008.
  5. de Jong, Ton, Marcia C. Linn and Zacharias C. Zacharia. “Physical and Virtual Laboratories in Science and Engineering Education.” Science 340, 305 (2013). doi: 10.1126/science.1230579
  6. Zane Education. “The Benefits of Using Educational Video in the Classroom.” Accessed 16 December 2015.
  7. University of Queensland Institute for Teaching and Learning Innovation. “Pedagogical benefits.” Accessed 16 December 2015.
  8. Robotów, Fabryka. “Robot programming languages.” Accessed 19 November 2015.
  9. Sanderson, Dan. “Alternative Programming Languages for Lego Mindstorms NXT.” 30 December 2009. Accessed 19 November 2015.
  10. Calin, Dragos. “Best Alternatives to Lego Mindstorms Kits.” 2013. Accessed 19 November 2015.
  11. Image taken from
  12. Image taken from
  13. Advances in Robotics. “International Conference on Robotics.” AIR in 2015 in Goa, India. 2-4 July 2015. Accessed 16 December 2015.
  14. Wood, Edward J.. “Problem-based learning.” Acta Biochimica Polonica 51, 2 (2004): XXI-XXVI. Accessed 19 November 2015.
  15. Weimer, Maryellen. “Problem-Based Learning: Benefits and Risks. Faculty Focus. 12 November 2009. Accessed 19 November 2015.
  16. Srinivasan, Malathi, Michael Wilkes, Frazier Stevenson, Thuan Nguyen and Stuart Slavin. “Comparing problem-based learning with case-based learning: Effects of a major curricular shift at two institutions”. Academic Medicine, 82, 1 (2007): 74-82. doi: 1097/01.ACM.0000249963.93776.aa
  17. Massachusetts Institute of Technology. “Robot 101: Learning to work with humans.” Robotics @ MIT. Accessed 19 November 2015.


Images on cover from:


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s