The ensuing project description was the inception to the academic adventure tasked with producing a controller that could regulate the suspension system on the CV90 MkIV platform with the use of a front looking, predictive, ground scanning sensor in addition to existing sensors. In other words, it was to create a pro-active suspension system for the BEA Systems Hägglunds’s made Combat Vehicle 90. The following report details the project of investigating, analysing and recreating the CV90 MkIV vehicle in theory and simulation. Thereafter it covers the implementation of two reinforcement learning algorithms, their corresponding training and subsequent results. Additionally the report discusses the possibility of a sensor technology with ground profiling properties as well as possible extension that the technology contains. The analysis of the vehicle used the newtonian approach whilst the recreation in simulation used the Python language as foundation. The simulation was then built by using the Python compatible MuJoCo physics engine. The MuJuCo model was constructed piecewise in a dedicated XML format, making it possible to quickly iterate over versions of the vehicle and environment. The reinforcement learning algorithms implemented were Policy Gradient and Proximal Policy Optimisation. The agents were then allowed complete control over the vehicle’s active suspension units whilst they learned by exploring their surroundings. The subsequent performance was then compared against a sky-hook control strategy implementation to determine their relative effectiveness. The proposed sensor technology was the ground penetrating radar system which has showed promise in a variety of fields and has the extended possibility of being used as an anti-tank mine detector. The project result showed that the feasibility of a reinforcement learning based controller is indeed possible, as it did achieve a marginal improvement in performance as compared to the baseline. However, further work is needed to reach a mature state since the final performance contained plenty of inefficiencies. Lastly there is an in depth discussion which provided a wide foundation on which future research could be built upon.