Course 'Design and Organisation of Autonomous Systems'

Project

• week 2: kick-off meeting (Monday 5th, 12:00, P.227)
• week 4: progress meeting (Tuesday 20th, 13:00, P.018)
• week 5: deadline draft article (Tuesday 27th, 16:00, pdf on website)
• week 5: deadline review article (Wednesday 28th, 16:00, see forms below)
• week 5: deadline final article (Thursday 29th, 16:00, pdf on website)
• week 5: mini-conference (Friday 30th, 14:00, F.013)

Groups

• AIBO Soccer - the Team
  Wiki site
  detailed assignment Arnoud Visser
  
  This year will be the 7th edition of the RoboCupSoccer - 4 Legged League. To facilitate progress in the League, the code is shared after the competition. To participate, on has to select the code with the best design, which makes it possible to extend easily it with our own algorithms. One of the most impressive participants of last year was the German Team, with many publications about their design. The task of the students is to evaluate their architecture.
  Final Report
  Review
• AIBO Soccer - the Pack
  Wiki site
  detailed assignment Peter van Lith
  
  Intelligence is about learning, learning is about playing. A four legged robot playing with a ball is an ultimate example of embodied intelligence, and the RoboCup is the ultimate challenge. The University that is playing with those robots the longest time is Carnegie Mellon. The task of the students is to evaluate their design, and to estimate how easy it will be to extend on their success.
  Final Report
  Review
• Self-reconfigurable distributed systems
  Wiki site
  detailed assignment Marinus Maris
  
  In factories for process control, control systems are used to monitor the ongoing production process. Such control systems typically comprise a large number of sensors and actuators. The sensors deliver information about the state of the control systems whereas actuators either regulate or can otherwise intervene with the process. However, the number of sensors and actuators (devices) in control systems is increasing steadily. Therefore the control of all these devices and hence the entire system is becoming very complex. This is a problem in case there are disturbances or even damages that require intervention from the operator, because the operator typically gets overloaded with error messages from the sensors and needs to select the appropriate action. Therefore, we have come to realize that complex control systems must be able to reconfigure itself, at least to some extend. This is called self-reconfiguration or, in case all components in the system are restored completely, self-healing. Hence, the question is how systems can react and reconfigure itself given certain observations from sensors.
  Final Report
  Review
• Abstract Java Camera project
  Wiki site
  detailed assignment Peter van Lith
  
  There is a growing interest in the application of small low-cost programmable camera systems for a variety of applications. Most camera systems depend on a frame grabber and software that runs on a PC. Developers of embedded applications are interested in using the same technology. Because the power of an embedded system is much lower than that of desktop or notebook PCs alternatives need to be found to allow a fair amount of processing power un such small systems.

  The Java Camera project aims at developing a direct interface between an standard CMOS camera module and a Java based hardware chip to allow real-time image processing to be done in standard Java. The development environment of this system also allows standard WebCam cameras to be programmed in Java, using the same architecture. The design of the Java Camera is inspired by the popular CMUCam project, in which a small microcontroller allows real-time object tracking for small robotic systems. This system however is not programmable and executes only the standard application. The Java Camera has a similar setup but allows full programming of the image analysis.

  In this project students will develop image analysis software, based on the Java Camera architecture and can choose from the following subjects:

  1. Motion Detection
  2. Object Counting
  3. Object Tracking

  Development will be done in Java on standard PCs and may be tested separately on the dedicated Java Camera hardware.

  Final Report
  Review