Mandli Communications recently participated in a LiDAR data collection project for the University of WisconsinMadison. The University currently runs a full-scale advanced driving simulator as a joint venture between the Traffic Operations and Safety Lab and the Cognitive Systems Engineering Lab. Mandli was able to provide LiDAR data that was utilized to simulate a real-life driving environment.
The driving simulator has been in operation since 2010, and is located in the Mechanical Engineering Building on the Madison campus. A full-sized Ford Fusion is installed on a motion platform within the simulator lab. The platform has the ability to provide the driver with the feeling of acceleration and deceleration within the vehicle. Through the windshield the driver is able to view the simulated environment, which is displayed on a high-resolution 240-degree cylindrical screen. The side mirrors feature displays as well, and a screen behind the vehicle is visible in the rearview mirror to complete the simulation.
As the driver operates the vehicle through accelerating, turning, and breaking, the simulated roadway on the displays responds appropriately. The displays are capable of updating the images consisting of the drivers point-of-view 60 times per second. The simulator also features a surround sound system, which is able to provide realistic driving sounds such as road, vehicle, and wind noise.
The simulation itself is powered by eight custom servers which handle the processing of the simulated environment and project the images, all while adjusting to any vehicle motion initiated by the driver. The environment is created by the three separate software packages, which handle the input from the vehicle, the behavior of the vehicle based on the desired simulation, and the construction of the virtual environment. Feedback from the simulator can be collected for a wide variety of factors, such as speed of acceleration and deceleration during travel, angle of the steering wheel, use of vehicle signals, etc. The simulated video is also recorded for later review and analysis.
In order to recreate a real-life environment, Mandli utilized our mobile data collection system to capture LiDAR data of an area in downtown Madison adjacent to the University campus. Our LiDAR system is vehicle-mounted, which worked very well in the recreation of an environment from the drivers point-of-view. For data collection we utilized a single HDL-64E LiDAR sensor from Velodyne. The sensor is encased in a rotating enclosure, and as it spins it collects over one million points of data per second. The unit is mounted onto the exterior of the collection vehicle which can safety collect data at highway speeds and within the flow of traffic, decreasing the risk to the traveling public and increasing the efficiency of the collection process.
Along with the LiDAR sensor a robust position system is also mounted on the vehicle, which includes a wheel encoder distance measurement unit. The positional system provides locational data that is synchronized with the LiDAR sensor, providing each point of the data in the LiDAR point cloud with its own unique positional coordinate. The vehicle also features a right-of-way imaging system, which was used for visual verification during the modeling process.
To create three-dimensional models from collected LiDAR data, Mandli imports the data into Autodesk Maya, a popular three-dimensional modeling program. In order for Maya to be able toread the LIDAR data Mandli created a custom software application to interpolate the raw data into a compatible format. Using Maya, a polygonmesh model is created of each section of the environment based off of the LIDAR data. Any gaps in the LiDAR data are manually constructed using visualization from the imaging taken during collection. The sections were then joined together to create a full-scale model of the environment.
UW-Madison was able to take the data collected by Mandli and import it into their software, creating a simulation that was very realistic in its representation of the real world. We are always interested in working with new technology, especially when it has the potential to create safer and more efficient vehicles and roadways. The driving simulator makes it possibles to test new vehicle technology and road infrastructures quickly and efficiently, and we are excited that we were able to contribute a small part to this important project.