In situ slip estimation for mobile robots in outdoor environments

Accounting for wheel–terrain interaction is crucial navigation and traction control of mobile robots in outdoor environments rough terrains. Wheel slip one the surface hazards that needs to be detected mitigate against risk losing robot's controllability or mission failure occurring. The open problems Terramechanics field addressed are (1) need situ wheel-slippage estimation harsh using low-cost/power easy integrate sensors, (2) removing prior information soil, which not always available. This paper presents a novel method utilizes only two proprioceptive sensors (IMU wheel encoder) estimate deep learning methods. It experimentally shown real-world feasible outdoor, uneven terrains without soil assumptions. Comparison with previously used machine algorithms continuous discrete show more than 9% 14% improvement performance, respectively.