Abstract
The prediction of vehicle mobility is an important aspect of vehicle locomotion system design. Mathematical models have been under development for many decades that estimate the tractive capability of large vehicles such as tanks and trucks across off-road terrain. These models have proven to be strongly successful when applied to these larger vehicles, but tend to vary in accuracy when applied to smaller vehicles. In such cases, soil trafficability experiments are often performed to determine whether or not a vehicle can traverse a given type of terrain. Unfortunately, this method of vehicle testing has limited applicability for planetary exploration vehicles, because planetary soils are almost non-existent on Earth. Various planetary soil simulants have been developed to replicate some of the mechanical properties of planetary soils, but the success of these in practice varies significantly. A significant limiting factor of traditional planetary vehicles is their inability to traverse the complicated terrain (such as rocks, steep slopes, etc.) of planetary surfaces. Future missions will need to consider more advanced locomotion systems beyond wheels and tracks. Unfortunately, trafficability models for these advanced systems either do not work effectively or do not exist at all. The latter is the case for legged vehicles. This thesis contributes to the understanding of microrover locomotion on planetary soils in various ways. Primarily, an accurate prediction model of legged vehicle mobility on the Martian surface is developed. In order to do so, two soil simulants are proposed and the mechanical properties of each are tested. Significant variation in these properties are exhibited due to varying levels of compaction and varying normal forces, each of which has an impact on the mechanical soil properties and the performance prediction of any vehicle type: wheeled, tracked or legged. In validating these soil properties, non-linearity was conclusively found in the critical shearing strength and the shear deformation modulus for a compressible soil, especially at low normal pressure, thereby challenging the globally accepted Mohr-Coulomb linear' relationship. The validation of a mathematical model to accurately predict the soil forces available to provide forward motion to a legged vehicle and the non-linearity of slip sinkage of a single leg in soil ai'e both conclusively validated experimentally. Simulations are then developed in MATLAB and with the Open Dynamics Engine physics library to evaluate the overall performance of a legged vehicle in Earth-based and planetary soils. Finally, a legged vehicle for planetary exploration is proposed, including the mobility prediction for the vehicle at various locations on the Martian surface. Keywords: legged locomotion, soil trafficability, vehicle mobility, planetaiy exploration, soil simulant, rough terrain. Legged Performance and Traction Prediction Tool, CAPTAIN rover.