Beijing University of Aeronautics and Astronautics (Beihang) and Huawei have jointly applied for a patent for "an uncertainty-based method for quantifying potential risks in vehicle blind spots".
According to the patent abstract, the invention discloses a method for quantifying potential risks of vehicle blind spots based on uncertainty, including steps:
Step 1: Obtain the original data of self-driving and observable traffic participants;
Step 2: Process the original data of the self-driving vehicle and observable traffic participants to obtain risk weight values, including: weight value of information uncertainty, weight value of blind area uncertainty, weight value of abnormal acceleration and deceleration of observable traffic participants, weight value of road network traffic flow in the autonomous vehicle driving area;
Step 3: Calculate the potential risk value of the blind area according to the risk weight value;
Step 4: Perform visual risk warning according to the potential risk value of the blind area.
This method does not need to add new sensors to realize the traditional perspective blind repair and reduce the cost. Based on the information uncertainty of the blind area, the abnormal acceleration and deceleration of the surrounding vehicles and the traffic flow in the road network, the danger degree of the surrounding environment can be predicted, which greatly improves the safety in the driving process.
Normally, the blind area is divided into the blind area inside the car and the blind area outside the car, which the blind area inside the car mainly has four visual blind areas: front blind area, rear blind area, rearview mirror blind area, AB column blind area. Blind spots outside the car are caused by fixed or moving objects and light problems.