Driving Simulator Test Bed
Organizer: Petros Ioannou
Sponsored: Fall 2012 – present
The MHI Driving Simulator (DriveSafety DS-100 ) is located in the Center for Advanced Transportation Technologies Laboratory in EEB305.
The below activity update for the driving simulator project was submitted by Petros Ioannou on 4/11/2013.
It is currently used to support the following activities:
Activity 1. Safety Evaluation of Advanced Driver Attention Systems (ADAS): The analysis of ADAS using an actual vehicle could be costly and dangerous as subject drivers cannot be subjected to dangerous driving situations. As a result the use of driving simulators to assess the effectiveness of these systems is crucial. In this activity we are developing vehicle following and lane changing scenarios where different Drivers will operate in conjunction with our version of ADAS which will provide warnings regarding unsafe situations as well as directions when to change lanes or merge into traffic. The driver’s response will be assessed for compliance and safety and used to modify and improve ADAS.
It is important to test the specific driving simulator’s validity before assuming it to be comparable to reality. In order to validate the data obtained using the driving simulator, we first compared the simulator to the real test vehicle to determine the simulator’s ability to provide a realistic driving environment. We plan to perform a series of speed and distance perception in real car and simulator comparison test. In addition to that we plan to compare vehicle following and lane change models corresponding to the same driver in a real car and the simulator in order to evaluate possible response differences.
To perform the above test we need to code the traffic environment the Drivers are going to operate. So far we developed a simulator map that resembles a district near the downtown Los Angeles and developed a methodology for testing the driving simulator versus the real car Driver perception.
Activity 2. Truck Driving in a Platoon: Under this activity we plan to develop a driving environment, which will be used to perform human factors tests associated with trucks in a platoon. In such situations we study how truck drivers respond to joining and exiting a platoon. We will start with manually driven trucks and build on to trucks that can be in the automated mode. In such case the driver response in switching between the manual and automated mode will be studied using the simulator.
Project Proposal approved 11/1/2012:
The driving simulator project will enable USC researchers to carry out various studies that include human factors, vehicle dynamics and interaction with driver, impact of sensors and warnings on driver, impact of different control techniques and automation on driver, driver assistance systems, information systems etc. The simulator can also be used as the starting point to improve its capabilities by developing additional driving scenarios and visualizations so there is also an element of computer science. The driving simulator will be used for research, individual directed research studies and projects. It will be accessible to all members of the EE Department and with permission on a case by case to students and researchers outside the Department. The simulator will be installed in the CATT laboratory at EEB 305.
After some research we selected the DriveSafety DS-100 simulator also used by other institutions and automotive research centers as the most appropriate for our possible research needs. The simulator is a research simulator that has three video channels (monitors) with 110° view angle, force feedback steering wheel, gear shifter, gas and brake pedals. The simulator allows us to create custom driving scenarios leveraging an extensive library of roads, intersections, cultural surroundings, traffic vehicles, pedestrians and other entities along with flexible scripting capabilities. In addition to the library custom 3D object could be imported to the simulator. The system includes powerful real time data collection and performance monitoring capabilities. Driving scenarios include basic autonomous traffic as well as custom-defined scripted vehicle actions and reactions. A broad research community facilitates the sharing of driving scenarios between industry experts. Authors c an control traffic signals, ambient traffic, scripted traffic, roadway friction, weather conditions, etc. Through the use of triggers, virtually any scenario can be designed.
The driving simulator has the ability to simulate 256 totally autonomous interactive ambient vehicles in the simulated environment. Vehicles are created, destroyed and controlled by the system with no scripting or user input required. These vehicles obey traffic laws, signage and signal devices. They interact realistically with other vehicles based on human behavior/decision models and real-time physics-based vehicle dynamics calculations. If specific behavior is desired, vehicles and other entities can be issued script commands through the use of triggers, timers, paths, routes and many other scenario tools.