BlueBotics (Switzerland)

ContentsUntil recently, autonomous mobile robots were mostly designed to run in indoor environments that are partly structured and flat. Many problems arise in rough terrain and position tracking is more difficult. Although odometry is widely used indoors (2D), its application is limited in natural environments (3D). The wheels are more likely to slip because of the rough structure of the soil, and the error in the position estimation can grow quickly. For these reasons, odometry is generally avoided in challenging terrains. However, we can look at the problem differently and ask: “How could we limit wheel slip in rough terrain ?”Two different aspects can be improved to minimize wheel slip. The first one is to develop a controller that guarantees good balance of wheel torques and speeds to optimize the robot’s motion. A torque controller minimizing slip and maximizing traction is presented in Chapter 4. The second aspect is to improve the mechanical structure of the robot. Indeed, a well-designed mechanical structure yields a smooth trajectory and thus enables limited wheel slip. As described in Chapter 2, SOLERO can adapt passively to a large range of obstacles and allows limited wheel slip in comparison with rigid structures such as four-wheel drive rovers. Thus, the odometric information is usable even in rough terrain. A new technique called 3D-Odometry, which provides 3D motion estimates of SOLERO, is presented in this chapter.

Abstract The design of a rover for a specific environment is a complex procedure which requires modeling a chassis and evaluating it with specific criteria. This is the aim of the performance optimization tool (POT) presented in this paper. The POT enables the comparison and optimization of a rover chassis in a quick and efficient way. The tool is based on a static approach including optimization of the wheel torques in order to maximize traction. Tests with real hardware were performed to validate the POT. Two different rovers, CRAB and RCL‐E, were assessed in simulation and hardware with respect to specific, well defined metrics. In simulation, their performances were compared to the rocker‐bogie‐type rover MER. CRAB and MER showed similar performance, while RCL‐E had significant problems with the benchmark obstacle. A very good match between simulation results and real measurements was achieved. © 2007 Wiley Periodicals, Inc.