Automated Design of Lightweight Exploration Rovers

In the recent years, vehicle-environment modeling techniques and powerful simulation tools have been used exhaustively to design wheeled rovers. In spite of that, rover preliminary design is still very dependent on human designer. It is also wellknown that human analysis of a complex vehicle dynamics is very time consuming, which implies in a simplified analysis of just a few useful operating conditions. It compels strict achievement of requirements without a deeper investigation of performance optimization potential during preliminary design phase. Our in-house developed rover optimization tool allow us to achieve a reasonable configuration having mobility and locomotion requirements and a given suspension concept as inputs. It reduced drastically the time usually devoted to synthesize some rover parametric configuration. We show the result of an optimized ExoMars rover under our scenario-oriented multi-objective optimization concept. The results are assessed through parameter variation studies to evaluate: allowable volume to place the center of mass of the vehicle, sensitivity analysis, Pareto frontier relating important metrics two by two, and figures of merit illustrating mapping of the design parameters into the criteria space. This research generates two branches of special interest: applicability of the current results (other than straight forward construction of the obtained suspension); and further development of the optimization tool. The current results can be applied to the selection of candidate actuator designs for example. It can be done by observation of the torque profiles of the vehicle's wheels as the vehicle drives on complex terrain in a simulation. This is used to evaluate candidate actuator design concepts which make the vehicle faster but at the same time do not enforce it to underperform in critical scenarios. Currently, our tool is able to perform geometric configuration optimization. But the results have shown that performance improvement is still possible according to the design of wheels, bogies and their mechanical connections (suspension). Optimization of locomotion components individually (e.g. topology optimization) and suspension concept synthesis are also discussed as means to enhance performance. Strategies still have to be investigated in order to define connection constraints in a suspension concept synthesis. This work summarizes results of a four-year research, but it can also be considered as the beginning of a second major step in the development of a fully automatized scenario-oriented tool to achieve complete rover synthesis including suspension concept synthesis, geometric configuration optimization, and component optimization.