Bio-Inspired Robot Design (Fall 2006)
Prerequisites: Permission of instructor.
Class Time: Tuesday and Thursday 1230 to 1400
Class Location: UC 153
Office Hours: TBD
Course Instructor: William C. Regli, Ph.D.
Instructor Contact: Email regli (at) drexel (DOT) edu (not that this will help any); Phone 215.895.6827; Office: UC 143
NOTE: This course is open to seniors and graduate students in any department in the College of Engineering (i.e. Computer Science, Mechanical Engineering, Electrical and Computer Engineering, etc). Space is limited, permission of instructor is required.
This course is being offered as part of a groundbreaking program sponsored by the National Science Foundation’s (NSF) Office of Cyber-Infrastructure (OCI) through 2008 (SCI-0537370 and OCI-0636235). Educators at Drexel University, The University of Maryland at College Park, The University of North Carolina at Chapel Hill and The University of Wisconsin at Madison have teamed to develop a set of unique multi-disciplinary course offerings on “Engineering Informatics” based around the exciting domain of “Biologically Inspired Robotic Systems”. This course in the first offering of this program at Drexel University
Engineers and computer scientists have discovered that the natural world can be successfully exploited to create novel software and engineered artifacts. This course specifically focuses on systems that are realized as physically embodied agents. Specifically, over the last several years, engineers have come up with new robot designs based on biological entities. These new designs offer significant benefits over the traditional robot designs. This new course will cover the fundamentals and applications of biologically inspired robots, with a specific focus on two important aspects of such systems:
- The tools and techniques for specifying robot behavior and representing knowledge about their design, function and behavior;
- The algorithms and software for supporting robot design, simulation, locomotion, control and coordination.
Topics to be Covered
- Fundamentals of Traditional Robotics: The history and taxonomy of traditional robots; different popular robot configurations; forward kinematics, inverse kinematics, and dynamics of serial manipulators.
- The Challenge and Opportunity for Bio-Inspired Robots: Bio-inspired robots are a new frontier in autonomous systems. Important areas, terrain, and applications that are not suitable for traditional robotic systems may prove suitable for bio-inspired robots. For example, search and rescue missions in complex urban environments require devices that can maneuver in collapsed buildings (e.g., cockroaches, ants), ductwork (e.g., spider, gecco), and other obstacles (e.g., lizard, centipede, etc). A snake-like robot may be used for planetary surface exploration, minimally invasive surgery, or inspection of piping and cabling. Such robots also have applications in homeland security and defense, enabling inspection of ships, containers and other structures too cramped for traditional robotic systems.
There are a multitude of challenges for these devices. Designing bio-inspired robots is a highly multi-disciplinary activity. Components include sensors, actuators, structural components, electronics, power source, and software. The actuated joints can create an extraordinary number of degrees of freedom (DOF). Joints must operate individually or be coordinated centrally; motion-planning algorithms must be rethought; and physics-based modeling and simulation tools need to scale to handle vastly greater complexity. For example, design and simulation of a snake inspired robot leads to numerous problems: these devices can include thousands of components, nearly all of which must interact for the snake to work. Additionally, one must capture engineering phenomena across all disciplines: mechanical, electrical, chemical, computer science, electronics, and environmental.
- Fundamentals of Biologically Inspired Robotics: This part of the course will begin with a discussion on the role of biological inspiration in robot design. Some of the questions being explored include “What can nature offer to engineers?” and “Can biologically inspired designs outperform traditional technology?” The next matter that is discussed is how engineers can quantify and evaluate nature in order to select the animal that best meets a set of design requirements. This part will discuss the maneuverability of animals and their ability to navigate various terrains. Several examples of bio-inspired robots will be discussed in detail, including the motivation and biological inspiration for their design, as well as technical specifications and comparisons to conventional robots. The examples will include robots inspired by the cockroach, snake, and tuna.
- Robot Informatics: The goal of this unit will be to help engineering students understand the issues involved in representing robot behavior (both aggregate and component) and design data. Some of the complex semantic relationships inherent in bio-robotic design include: (a) Knowledge representation of the domain semantics for bio-robotic systems; (b) Introduction to glossaries, thesauri, dictionaries, ontologies, and taxonomies; (c) What are the uses for ontologies, especially in the context of engineering?; (d) Abstract methods and formalisms for understanding, representing, and communicating essential properties of domain-specific engineering systems. This includes being able to write formal specifications using domain independent modeling languages; (e) and system integration principles using formal logic and semantic web technology (OWL, OWL-S), as well as use formal modeling tools such as UML.
- Systems Engineering & Multi-Disciplinary Design: Design is a struggle to teach in Universities. The main technique is a project class, the result of which is the immersion of students into a thicket of human and system interactions in which they learn design principles by doing. Even then, design is rarely taught from a multi-disciplinary, system engineering, view. For example, the interplay between software and physical system has been the downfall (e.g., Denver Int'l Airport baggage claim) or near downfall (e.g., F-22 Raptor) of many a system. This is especially prevalent in systems where software is a major component and the overall system designers are not versed in informatics principles. A key challenge is partitioning a complex design problem into individual subsystems. This module will present students with a multi-disciplinary view of the design. They will be introduced to the overview of each relevant discipline (i.e., software spirals, waterfall model, mechanical design, electronics, energy systems), as well as several general systems engineering principles (i.e., systems modeling, system performance assessment, trade-off analysis, system partitioning, etc). Co-design of physical components and software will be explored in detail.
The above topics could easily fill 2-to-3 academic quarters. The goal of this class is not to replicate subjects that are already covered in depth elsewhere, rather to present an integrative view of how they must mesh to properly design and control bio-inspired robotic systems.
Lectures will cover, at varying degrees of depth, all of the topics above. In several case, remote lectures by our partner institutions will be used to augment materials presented at Drexel. Specifically,
|University of Maryland||Overview of Bioinspired Robotics|
|University of North Carolina||Kinematics, Simulation and Physics Based Modeling|
|University of Wisconsin||Dynamics, Mechanical design, CAD, rigid body modeling|
We expect to have six (6) lectures during the course of the quarter delivered via remote video or other means by these other institutions. Project Based Learning: Robot Building Projects This course will emphasize hands-on learning. As a part of the course projects, student teams will have an opportunity to design, model, simulate and (hopefully) build their own snake-inspired or bio-inspired robots. These projects are expected to provide a very valuable experience for students of all disciplines.
Week by Week Breakdown of Lectures
|Date||Topic||Links and References|
There is no formal text for this class. Course lectures will be augmented with reading materials, technical papers and web materials.
Interested students may wish to examine the following textbooks:
- J.J. Craig. Introduction to Robotics: Mechanics and Control. Prentice Hall; 3rd edition, 2003.
- Kiyoshi Toko. Biomimetic Sensor Technology. Cambridge University Press, 2000.
- Karl Williams. Amphibionics: Build Your Own Biologically Inspired Reptilian Robot. McGraw-Hill/TAB Electronics, 2003.
- David Cook. Robot Building for Beginners. Apress, 2002.
Students will gain experience with a variety of computational tools, including MATLAB, SolidWorks, Pro/ENGINEER, ACIS, MAPLE, etc. Several research prototype tools may also be introduced for physics-based modeling, kinematics, etc. Assignments and Evaluation Criteria
Course Objectives and Outcomes
The goal of this class is to build comprehensive engineering models of biologically-inspired robotic systems. Students successfully completing this class will
- be able to identify problems resulting from the interdisciplinary interactions in bio-inspired robots;
- perform system engineering to design, test and build bio-bots;
- be able to apply informatics principles to bio-bot design and testing;
- gain experience using a variety of pedagogically appropriate hardware (i.e. Lego Mindstorms, Roombas, etc) and software tools (see above) for robot design/analysis.
Assignments and Grading
All assignments in this class will be turned in over the course Wiki. All students will be expected to get accounts on the Wiki and use it to document weekly progress, as well as author final reports:
- Weekly Progress Reports (in Wiki) 50%
- Final Project Report (in Wiki) 50%
There is no final exam; the final project will be due in lieu of a final exam.