Autonomous driving and driver assistance systems require accurate information about the vehicle and its surroundings to perform tasks such as robust path planning. An occupancy grid map can provide such information, but it too requires precise information about the vehicle’s location. We present an approach to Monte Carlo Localization on an occupancy grid map based on an automotive lidar sensor cocoon providing 360° measurements around the vehicle using five Valeo SCALA sensors. Standard MCL is enhanced through an alternative particle weighting function and separate alpha filters are used to incorporate odometry measurements. Additionally, scan point sampling is introduced into the particle weighting function to select scan points most representative of pose estimation quality. Compared to paths reconstructed from only the vehicle’s odometry signals, the mean squared error in heading angle and position is reduced by 93-97% and 86-96%, respectively. Investigated scenarios include urban roads, factory roads, elevated country roads and highways.