The SmartTer - a Vehicle for Fully Autonomous Navigation and Mapping in Outdoor Environments

The driving factor for the development of the vehicle presented in this paper was to construct a hardware platform that allows to perform the tasks of environment mapping and autonomous navigation in large scale outdoor environments. Our robot is based on a standard Smart car that has been equipped with five distance laser sensors, three cameras, a differential GPS, an Inertial Measurement Unit (IMU), an optical gyroscope and four computers. The car’s systems states are directly accessed through the vehicle’s CAN bus. Localization and Navigation are realized by fusing all available sensory information in a probabilistic way. This allows for high precision localization and dynamic local and global path planning. Using the 3D point clouds extracted from the rotating lasers and the omnidirectional image the smartTer is able to consistently register the 3D maps and analyse the scene for regions of interest. To control the car in both longitudinal and lateral direction we applied a couple of modifications to the car, namely providing access to the car’s power steering system, electronic gas and brake pedal. Keywords—sensor fusion, outdoor vehicle, 3D mapping