Within the framework of the National Lab for Autonomous Systems (ARNL) we made demonstrations for C-ITS (Cooperative Intelligent Transport Systems) services for autonomous / highly automated vehicles and intelligent traffic lights.
The demonstration includes a self-driving go-kart developed at the department and PLC-controlled traffic lights. The go-kart does wireless communication to a central computer (quasi traffic control center) that also controls the traffic lights. In addition, a digital twin of the demonstration runs in real-time in a 3D simulation environment.
Communication is a sophisticated way to share information and make decisions based on that knowledge. Extending intelligent vehicle functions with communication raises the safety and performance of transportation and opens the way towards fully autonomous driving which is engineers’ primary objective. Avoiding accidents and protecting human lives while creating better comfort and efficient trips with passenger cars requires smart solutions.
A newly designed, high-performance algorithm was created and tested in a complex virtual environment. While today’s simulation techniques are getting more and more reliable, developing the actual algorithm on hardware is always challenging. With the experiences from the simulations, the algorithm was carried out and tested in real circumstances using professional communication hardware from the automotive industry.
The algorithm can be used in two ways. It can be used as a driver assistance system for regular human-driven vehicles; however, it can also control fully autonomous vehicles. The algorithm was designed to control incoming vehicles at an intersection in a distributed fashion, meaning there is no need to deploy traffic lights or RSUs (roadside units). It makes the algorithm fail-safe as it does not only depends on single hardware but each vehicle is considered in the decision process.
The control algorithm’s procedure is straightforward. It works based on the “right-of-the-way” rule for non-signalized intersections. It was designed to maintain safety even if only one single V2X message arrives from the arriving vehicle, making it robust against communication interference and delays. The algorithm was demonstrated in a “T” shape intersection, assuming the right-of-way rule is applicable on the spot. Two regular human-driven vehicles were taking part in the demonstration. The algorithm worked as a driver’s assistance system in this demonstration. The algorithm was connected to an HMI (Human-Machine Interface). In real-time, this HMI visualizes vehicles around the user based on standardized automotive communication messages called Cooperative Awareness Messages (CAM). The intersection control algorithm relies only on data from these CAM messages. As a driver approaches the junction, the algorithm decides about the priority if there are incoming messages from other vehicles. The HMI shows various signals to help the drivers decide if they can move freely or have to give priority to the others.
It is also possible to focus the camera on the incoming vehicle to monitor the other vehicle’s locomotion. The key function and reason for the performance gain are that with the help of this algorithm, drivers do not have to slow down even at intersections, where the environment blocks the driver from seeing if there are other arriving vehicles with priority. Signals consider various plausibility checks (such as the time between two messages and the working connection between the V2X hardware and the HMI). The system will raise the driver’s caution if any fault is suspected. The visualization system is available for multiple platforms (such as iOS, iPadOS, and Android). It was tested with tablets and a smartphone for demonstration purposes. The HMI also provides a settings interface to fine-tune the algorithm. The demonstration was carried out by testing all possible variations of the right-of-the-way rule, which was successful.