Qualitative Relational Navigation using Minimal Sensing (NSF)
The goal of this research is to develop a navigation methodology for robots using relational geometric models that are constructed and evaluated using minimal sensing.
1. Sensing, Representations and Map Formulation: we formally define qualitative sensing, along with the underlying mathematical representations and operators to develop qualitative relative maps.
2. Considering Uncertainties: we expand sensing and mapping to include probabilistic uncertainties in the sensor likelihoods and estimation approaches.
3. Navigation: we develop a sampling based approach to navigation that intimately builds on the qualitative state representation.
4. Experimental Validation: we seek to evaluate the maturity of the proposed algorithms via an increasingly complex set of experimental validation tests using two testbeds in the PI’s lab.