Pebody, Miles;
(1997)
Autonomy in the real real-world: A behaviour based view of autonomous systems control in an industrial product inspection system.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
out.pdf Download (24MB) | Preview |
Abstract
The thesis presented in this dissertation appears in two sequential parts that arose from an exploration of the use of Behaviour Based Artificial Intelligence (BBAI) techniques in a domain outside that of robotics, where BBAI is most frequently used. The work details a real-world physical implementation of the control and interactions of an industrial product inspection system from a BBAI perspective. It concentrates particularly on the control of a number of active laser scanning sensor systems (each a subsystem of a larger main inspection system), using a subsumption architecture. This industrial implementation is in itself a new direction for BBAI control and an important aspect of this thesis. However, the work has also led on to the development of a number of key ideas which contribute to the field of BBAI in general. The second part of the thesis concerns the nature of physical and temporal constraints on a distributed control system and the desirability of utilising mechanisms to provide continuous, low-level learning and adaptation of domain knowledge on a sub-behavioural basis. Techniques used include artificial neural networks and hill-climbing state-space search algorithms. Discussion is supported with examples from experiments with the laser scanning inspection system. Encouraging results suggest that concerted design effort at this low level of activity will benefit the whole system in terms of behavioural robustness and reliability. Relevant aspects of the design process that should be of value in similar real-world projects are identified and emphasised. These issues are particularly important in providing a firm foundation for artificial intelligence based control systems.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Autonomy in the real real-world: A behaviour based view of autonomous systems control in an industrial product inspection system |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Thesis digitised by ProQuest. |
| Keywords: | Applied sciences; Autonomous systems control; Autonomy; Industrial product inspection; Real-world world; System |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10097817 |
Archive Staff Only
 |
View Item |
