Keywords autonomous robots, knowledge-based systems, mobile robots, navigation, perception
Start Date: 21-DEC-88 / Duration: 48 months
[ contact / participants ]
The objective of this project is to develop integrated hardware and software systems to give autonomy to a class of robot vehicles and manipulators in an environment, parts of which can change dynamically.
A key issue within the project is the development and testing of an information framework called the Perception and Control Model. Based on this model, the decision system uses a combination of symbolic and numeric techniques, thus combining the strong points of both control engineering and artificial intelligence approaches. The combination of the two approaches will be demonstrated in two distinct testbeds. The first emphasises path-finding and coarse manoeuvring capabilities in a robot vehicle, the second highlights coarse and fine manoeuvres as well as flexible mission planning, as required in a diagnostic data acquisition task.
The integration of symbolic and numeric models will enable these systems to meet the major requirements of the control system:
Currently, the development of the necessary hardware is mostly completed, and design, implementation and testing of the software is in progress. A major result of the project is the new concept of an "exception based" architecture where classical and sensor based control are combined, e.g. in the use of logical sensors, sensor based control, fuzzy control, and non-linear filtering techniques.
The techniques and components developed by the project are foreseen to have applications in mobile robots used within CIM installations and in hostile environments, e.g. the storage of nuclear waste products or other poisonous, inflammable or explosive materials, and the dismantling of nuclear power-plants. Automotive applications are also anticipated, for example in subsystems assisting the driver in dangerous situations or in taking over tedious driving tasks.
The industrial challenge of both test-beds is to perform complex missions in an autonomous and reliable way, using a combination of sensors to realise a robust system with affordable components.
Mr Steef Geenen
CAP VOLMAC BV
NL - 3500 GN UTRECHT
tel: + 31/ 30-526286
fax: + 31/ 30-546451
telex: 70644 VOLM UT
CAP VOLMAC BV - NL - C
CSIC - E - P
FRAMATOME GROUP - F - P
HITEC LTD - GR - P
ROBERT BOSCH GMBH - D - P
UNIVERSITY OF STRATHCLYDE - UK - P
INDECON - GR - P
FRAMENTEC SA - F - P
UNIVERSITEIT VAN AMSTERDAM - NL - P
MARIE - 2043, December 1993
please address enquiries to the ESPRIT Information Desk
html version of synopsis by Nick Cook