This study investigates the computational speed and accuracy of two numerical integration methods, cubature and sampling-based, for integrating an integrand over a 2D polygon. Using a group of rovers searching the Martian surface with a limited sensor footprint as a test bed, the relative error and computational time are compared as the area was sub-divided to improve accuracy in the sampling-based approach. The results show that the sampling-based approach exhibits a 14:75% deviation in relative error compared to cubature when it matches the computational performance at 100%. Furthermore, achieving a relative error below 1% necessitates a 10000% increase in relative time to calculate due to the $\mathcal{O}\left( {{N^2}} \right)$ complexity of the sampling-based method. It is concluded that for enhancing reinforcement learning capabilities and other high iteration algorithms, the cubature method is preferred over the sampling-based method.

One of the more complex challenges to be overcome when modelling the dynamics of any mobile platform is the representation of the forces of friction. In this work, a verified friction model for a four-wheel planetary rover with skid steering is set out. The model is implemented in multiple simulation scenarios to cover the full scope of the rover's mobility (forward motion, turning, braking, and accelerating) in both 2D and 3D environments.

Fault tolerant design is hugely important for autonomous mobile robots such as planetary exploration rovers (PERs), as they are required to be both robust and reliable in extremely harsh environments. One of the main principles of fault tolerance is the detection and diagnosis of any faults afflicting the system. A model-based fault detection procedure is presented using forward and inverse simulation methods. The results of each method are compared for faults in different system locations to display the differences and advantages of both methods. It is shown by this comparison that the methods complement each other and can be used concurrently to diagnose output and input faults.

A taxonomy of planetary exploration rovers is presented, followed by a review of systems used in missions and in an experimental phase. The baseline design emerges as four or six wheels, rocker-bogie based passive suspension and all wheel driving/selected wheel steering. A trend is also apparent in the use of wheel - legged hybrid locomotion. The performance metrics are presented by which the differing configurations of the locomotion subsystem for wheeled rovers with a passive suspension may be systematically evaluated. The taxonomy and aggregated metrics presented in this paper aid in the comparison and selection of rover characteristics, while the baseline design is a representative example of current practices and future trends.