Abstract
One of the most challenging problems in robotics is the design of navigators. Navigators are motion controllers that are used to generate a constrained solution trajectory linking a starting point to a target zone for an agent of arbitrary, unknown shape that is operating in an unknown environment. In other words, this class of controllers is required to carry out the unusual task of solving an ill-posed problem. This paper discusses the construction of a hybrid control structure that would enable an agent to successfully deal with the informationally-deprived situation that is described above. The control structure has an evolutionary nature that allows for Autonomous Structural Adaptation of the control relying only on a sequentially acquired sensory input as a source of information. It also has the ability to integrate any available a priori information in its database, regardless of its fragmentation or sparsity, in the planning process to accelerate convergence. The structure is based on a bottom-up, Artificial Life approach to behavior synthesis that allows the integration of experience along with synergy as drivers of the action selection process. Theoretical development of the structure, a realization, and simulation results are provided.
Original language | English |
---|---|
Pages (from-to) | 2535-2540 |
Number of pages | 6 |
Journal | Proceedings of the IEEE Conference on Decision and Control |
Volume | 3 |
State | Published - 1998 |
ASJC Scopus subject areas
- Control and Systems Engineering
- Modeling and Simulation
- Control and Optimization