Variable Object Identification, Location and Acquisition


VOILA - 2502

Keywords vision systems, autonomous vehicles


Start Date: 01-APR-89 / Duration: 36 months

[ contact / participants ]


Objectives and Approach

The aims of the VOILA project were to carry out a collaborative programme of research on the development of flexible, dynamic vision systems and to prove their feasibility for application to a range of tasks in the operation of robot vehicles in a variety of industrial and commercial environments. The objectives of the vision research carried out were to:

Meeting these objectives required significant advances in the state of the art: in early vision and the use of stereo and motion, in the use and exploitation of predictive feed-forward techniques and dynamic vision, in the range and type of models that can be utilised in a vision system, in the design of vision system architectures and their control structures, and in the implementation of vision systems on multiprocessor environments.

Progress and Results

A number of basic visual competences involving obstacle detection, free space determination, local navigation, global map-making, and object recognition, location and acquisition were developed and demonstrated on vision and vehicle experimental platforms at four laboratory sites within the project. Two of these were further developed to show the potential of passive vision for enhancing the capabilities of mobile vehicles in teleguided operation and in autonomous systems. A third final demonstration was used to illustrate how different vision systems and modules could be integrated on a multiprocessor distributed environment. The fourth platform was used to demonstrate the potential of one of the systems developed for operation in outdoor scenes.

Exploitation

Industrial evaluation has been an on-going activity in the project since it began. Industrial requirements for improving both the flexibility and autonomy of mobile robot systems were instrumental in framing the design of the experimental platforms and demonstrations. The performance of the systems developed was assessed and their potential for industrial application evaluated. In the final year of its operation, the consortium mounted a series of publicity and marketing actions including seminars, workshops and a project Open Day.

In addition, an analysis of prospective markets and applications was carried out. These included the areas of most interest to the industrial participants, GEC, ELSAG-Bailey, MS2i and Roke Manor Research: advanced robotics, including telerobotic operation of systems in hazardous environments such as nuclear power plants and off-shore; agricultural robotics, especially for adaptive operation of mobile robot systems; space robotics, for example for in orbital and planetary operations; and the computer vision and image-processing market, where high-performance systems are required.


CONTACT POINT

Prof Bernard Buxton
GEC MARCONI
Hirst Research Centre, East Lane
UK - WEMBLEY HA9 7PP
tel: + 44/ 81-908-92-81
fax: + 44/ 81-908-90-90
telex: 923429 GECLAB

Participants

GEC MARCONI LTD - UK - C
ELSAG BAILEY SPA - I - P
MATRA MS2I - F - P
UNIVERSITA DI GENOVA (DIF) - I - P
UNIVERSITY OF SHEFFIELD - UK - P
UNIVERSITY OF OXFORD - UK - P
ROKE MANOR RESEARCH LTD - UK - P
INRIA, SOPHIA ANTIPOLIS - F - P
GUIDANCE & CONTROL SYSTEMS LTD - UK - A
AITEK - UK - A


ST synopses home page ST acronym index ST number index
All synopses home page all acronyms index all numbers index

VOILA - 2502, December 1993


please address enquiries to the ESPRIT Information Desk

html version of synopsis by Nick Cook