Robinson, M., and Jenkin, M. Reactive control of a mobile robot, In C. Archibald and P. Kwok (Eds.), Research in Computer and Robot Vision, pp. 55-70, World Scientific Press: Singapore, 1995.
This paper describes the design and implementation of a robot guidance system which integrates a classical occupancy grid based global path planner to construct a global plan and low-level subsumption based architecture to safely execute it. The global path planner generates a correct path from the robots current state to a goal state according to a static map of the world, and low-level architecture is given the task of avoiding any unexpected or unmodelled obstacles. By dividing the task of driving the robot to a mapped goal into a classical AI task and a reactive subsumption one, the ARK robot takes advantage of the best technologies. Results obtained using a Real World Interface (RWI) B12 mobile base, an onboard reactive control system and the ARK global path planner are shown.
Dudek, G., Jenkin, M., Milios, E., and Wilkes, D., Experiments in sensing and communication for robot convoy navigation, Proc. IEEE/RSJ IROS, 268-273, 1995.
This paper deals with coordinating behaviour in a multi-autonomous robot system. When two or more autonomous robots must interact in order to accomplish some common goal, communication between the robots is essential. Different inter-robot communications strategies give rise to different overall system performance and reliability. After a brief consideration of some theoretical approaches to multiple robot collections, we present concrete implementations of different strategies for convoy-like behaviour. The convoy system is based around two RWI B12 mobile robots and uses only passive visual sensing for inter-robot communication. The issues related to different communication strategies are considered.
Jenkin, M. R. M., Milios, E. E., and Tsotsos, J. K., Cyclotorsion and the TRISH active stereo head, Proc. Int. Workshop of Sterescopic and Three DImensional Imaging, Santorini, Greece, 1995.
For many stereo applications fixed stereo geometry has been found to be inadequate and active stereo systems capable of manipulating the camera geometry have been developed. Head motions, and individual camera pan and tilt motions define the fixation poin of an active system, while rotations of the cameras about their optical axes (torsion) defines the local shlope of the zero disparity surface near the fixation point. As stereo processing typically prefers a surface slope perpendicular to the plane containing the fixation point and the nodal points of the two cameras, torsional control in an active stereo system such as TRISH can be used to improve the efficiency of the search for stereo matches and can tune the head geometry for specific object recognition tasks. This paper explores the relationship between head geometry, including non-zero camera torsions and the position and shape of the near-zero disparity surface. It develops a zero-disparity and zero-cyclo-disparity control system for tracking both the position and slant of a target using an active stereo head capable of torsional camera motions.
