A few days ago, some media learned from officials that Tesla has pushed the FSD (Fully Autonomous Driving) 10.6.1 Beta version. This version is a beta version. This update is mainly aimed at improving functions such as automatic braking and object detection network architecture, so as to achieve higher detection accuracy of vehicles on objects. This update brings a number of improvements, improves the accuracy and speed of object detection, and solves some of the annoying problems that existed before. It seems that Tesla is taking great pains to improve the driving experience of car owners.
This version improves the accuracy of monitoring VRU objects such as pedestrians, bicycles, and motorcycles by updating the next-generation detection algorithm. On the basis of the existing FSD Beta 10.5, the Tesla FSD 10.6.1 Beta version improves the recognition accuracy, reduces the recognition error at night and decrease the average relative error by 18.5%; at the same time, it makes concessions to oncoming vehicles and slows down corrections when lanes deviate. The stability of the vehicle is gradually approaching the assistance level of L5 automatic driving; when an object crosses the front lane, the automatic braking action is improved to make it smoother and enhance passenger comfort; and improve the parking and waiting position when the vehicle turns left at an open intersection. In addition, the Tesla 10.6.1 Beta version also optimizes the recognition accuracy, unprotected left turn parking position, and slows down the stability of lane departure.
Tesla CEO Elon Musk previously said, “Tesla will launch a fully automated driving (FSD) software package subscription service in the second quarter of 2021. The main functions of FSD include Autopilot navigation, automatic lane change, and automatic parking,summoning function, FSD computer, signal light and parking sign control. FSD one-time purchase cost is about 10,000 US dollars." Although Chinese users can also purchase Tesla's FSD, due to China's complex traffic environment and lack of high-precision maps that can be localized, it cannot be used on Chinese roads.
