10-08-2012, 03:31 PM
Introduction to Robotics
Intro2Bots.doc (Size: 39.5 KB / Downloads: 87)
Another educational initiative based on the use of microcontrollers is the newly developed Introduction to Robotics course. This course is geared toward junior/senior level undergraduates and beginning graduate students in computer science and artificial intelligence (the University of Georgia offers BS, MS, and PhD degree programs in computer science as well as a separate MS degree program in artificial intelligence). The primary focus of the course deals with all aspects of autonomous mobile robots. In particular, the major issues investigated are cognitive behavior, and motion. Cognitive behavior addresses problem solving using sensory inputs and desired goals. Motion deals with aspects of movement in the real world from simple fixed-base robotic arm movement to autonomous rovers in unknown environments.
Students completing the Introduction to Robotics class will have been exposed to a number of lecture topics as well as many practical (hands-on) topics. Lecture topics include introduction to robotics, cybernetics, history of robotics, robotics in fact and fiction, application areas, mechanical foundations, electrical foundations, control, intelligent behavior, autonomous robot architectures, robot reasoning, knowledge representation, and planning. Practical topics include robot kit construction, wiring diagrams, simple circuits and components, basic electronics, soldering, motors, gears, principles of motion, microcontrollers/microprocessors, sensors, feedback, and computer programming for intelligent behavior. The course consists of traditional lecture activities and hands-on laboratory activities.
Performance measurement within the class is based in part on traditional assignment/testing instruments. However, the majority of a student’s grade is based on laboratory exercises. Laboratory reports are prepared by teams of three to five students working to resolve a specific laboratory challenge. Each challenge is designed to achieve an educational objective involving mastery of various robotics related topics, problem solving creativity and innovation, team organization and management, and verbal and written presentation skills. At a more detailed level, each challenge provides students with ample opportunity to become immersed in the mechanical and behavioral aspects of the various robots used in the class. At a more abstract level, students are exposed to a simulated work environment (as close to a real-world environment as possible) where a team is given the task of achieving a specific goal within a specific time frame while using available tools and equipment. Performance is based on results not effort, on quality of team output not on the superiority of work done by any one individual. Since the challenges are team-based, a mechanism is in place to translate a team “grade” into an individual grade for each of the students on a team. This mechanism ensures that credit is given to those members of the team who earn and deserve it.
The “maze egress” challenge requires each robot to exit a maze in a short amount of time. By the time the class is ready for this assignment, the challenges have evolved into competitions among the teams. The challenge “winner” in this case is the team whose robot can exit the maze in the shortest time (maze configuration is random and unknown to the teams prior to egress demonstrations). The behaviors involved in maze egress include reactive behavior, selective random motion, and some small amount of learning. Typically, infrared range sensors and touch sensors are the primary sources of inputs to the robots for this challenge; however, some creative teams include interesting distance measurement schemes as well as some sort of terrain mapping memory scheme. The maze is constructed on a large conference table and has an outer boundary wall as well as internal passageways. The internal wall configuration disallows naïve maze egress schemes such as simply following the left-hand wall until an exit is reached. From a resource supply and resource usage point of view, students quickly learn which ideas work best for the various situations their robots encounter. For example, they learn to develop robot behaviors that “recognize” blind alleys that require the robots to back up rather than turn around, non-productive repetitive movements such as being “stuck” in an infinite reaction loop, and remembering previously visited passageways that did not lead to an exit.