Multi-sensor fusion is an inevitable part when designing an autonomous vehicle. The integration of measurements coming from different sensors has the advantage that the navigation system is more likely to provide a solution in a more varied range of scenarios. Due to the nature of the measurement principle of each sensor, the error dynamics are completely different and the strengths of a certain sensor compensate for the weaknesses of another. Nevertheless, in order to design and tune our navigation system according to the application we are targeting, it is needed to have a flexible framework where we can test different type of sensors, different sensor configurations, environments (urban, semi-urban, rural) and vehicle dynamics. Acquiring data from measurement campaigns is not only costly and time consuming, but usually one cannot have complete control of the environment and it is also challenging to acquire the ground truth data. For that reason, in this work, we present an extension to a previously introduced framework called SIMSS [1]. This framework is currently able to generate GNSS, IMU and LiDAR synthetic data and process it an RTK/INS/LiDAR loose coupled Kalman filter. In this work, we successfully show the quality of the generated synthetic data and the integration with our navigation modules within the MuSNAT. Hence demonstrating its use in testing and verification of our algorithms, which in turn gives us the opportunity to offer a more robust navigation system.    «

Multi-sensor fusion is an inevitable part when designing an autonomous vehicle. The integration of measurements coming from different sensors has the advantage that the navigation system is more likely to provide a solution in a more varied range of scenarios. Due to the nature of the measurement principle of each sensor, the error dynamics are completely different and the strengths of a certain sensor compensate for the weaknesses of another. Nevertheless, in order to design and tune our naviga...    »