Digital Twin for the MMX Rover Locomotion Subsystem

A crucial part of the operation of the MMX rover on the surface of Phobos will be an in-depth understanding of the behavior of the rover's locomotion subsystem. The low gravity, roughly 1/2000 of that on Earth, makes laboratory experiments difficult. Thus, simulations are indispensable for understanding or predicting the robotic behavior of the rover. This paper focuses on developing and validating the locomotion subsystem model. Building upon the concept shown in a previous publication [1], the modular approach in modeling is shown in detail with the example of the locomotion subsystem. Simulation models exist for the various components of the locomotion subsystem (LSS). They range from simple static components representing only geometry to complex mechanical and electronic elements. These are then combined into configured versions of the locomotion subsystem model for different applications. A main part of this paper is to present and motivate the different modeling techniques: physical, differential equations or look-up-tables. In addition to the electromechanical system, the behavior of the components in the electronic box (Ebox) is represented. The behavior of analog to digital conversion for sensor signals and the logic implemented on an FPGA is reproduced in simulation. This model includes an interface to connect the simulation to either a hardware-in-the-loop (HiL) setup used at CNES for system verification or a software-in-the-loop (SiL) setup used at DLR for locomotion software development and as a locomotion planning tool. The rover simulation must be representative of the actual flight hardware for application during operations. This matching is achieved by parametrizing the model to match the available measurements of the LSS. Various measurements from component and system levels are used to confirm the model's validity and justify the term “digital twin”.