According to foreign media reports, Tesla CEO Elon Musk said the company will introduce automatic identification of traffic lights and parking signs in the United States in the next few weeks.Tesla said in the function introduction that when automatic steering and traffic perception cruise control are activated at the same time, the car can automatically stop at the traffic lights and stop signs.
Both features are included in Tesla's basic version of autopilot, which is standard on all of the company's current vehicles.
Tesla's traffic lights and stop sign recognition have been planned for a long time. In March 2019, Tesla released a new automatic steering stop warning function. When the driver approaches the intersection, the system automatically reminds the driver to take over the operation of the electric vehicle.
The introduction of traffic light and stop sign recognition function and automatic stop function will further improve the safety of autopilot.
Tesla said that due to different traffic rules and related policies and regulations around the world, autopilot updates for customers in other countries will wait longer than for us owners.
Before that, more than 600000 Tesla cars were equipped with a fully automatic driving chip, which has a transistor of up to 6 billion, can complete 144 trillion calculations per second, and can process 2300 frames per second at the same time, and each car has two such chips, which can process the same data at the same time.
It then applied for a patent on how to obtain training data from its huge customer fleet to train its autonomous driving neural network.
How to Implement Automatic Driving
At present, the key technologies of automatic driving are perception, planning and execution, which involve sensor, data processing, machine learning, slam and sensor fusion, path planning and other fields.
Among them, perception is mainly responsible for the calculation of vehicle behavior through hardware interaction and communication of vehicle sensors, and control is the electronic operation of vehicle components.
perception is mainly positioned through environmental perception.
In the stage of environment perception, it is necessary to obtain a large amount of surrounding environment information to ensure the correct understanding and corresponding decision of the automatic vehicle to the surrounding environment of the body.
Environmental awareness is the ability to understand the environment. For example, Lane signs and markings, traffic light recognition, traffic signals and signs recognition, pedestrian and vehicle detection, types of obstacles and other data understanding, analysis and classification. Positioning is the post-processing of perception results. Through positioning function, it helps the automatic vehicle to understand its position relative to the environment.
In the planning part, according to the original data and existing maps captured by the autopilot vehicle sensor suite, the automatic driving system needs to build and update the specific environmental map through the simultaneous localization and mapping algorithm, track its specific location, and start planning the path from one point to another point.
At present, the latest development of machine learning is to effectively deal with the data generated by the sensor of an autopilot vehicle and reduce the computational cost. In addition, advances in chip fabrication and miniaturization are improving the computing power that can be installed on autonomous driving vehicles.
5g's high bandwidth and low delay will help the network-based data processing to operate autonomously.
Execution is that the system controls the vehicle according to the decision results.
Each control system of the vehicle needs to be able to connect with the decision system through the bus, and can accurately control the driving actions such as acceleration degree, braking degree, steering amplitude, light control and so on according to the bus instruction issued by the decision system, so as to realize the autonomous driving of the vehicle.
Perceptual positioning is like the driver's eyes, planning and decision-making is equivalent to the driver's brain, and executive control is like the driver's hands and feet. Executive control is the foundation of the real landing of automatic driving.
Tesla's fully automated driving program
The biggest selling point of Tesla electric vehicle is its automatic driving system (the current technology level belongs to the two class "auxiliary driving" level), and the company will upgrade its automatic driving system every once in a while.
The basic principle of Tesla's automatic driving system is to accurately identify objects and objects on the road, and the driving computer will make various driving control actions according to these identified road conditions.
Musk even plans to set up an unmanned taxi company to let a large number of self driving electric vehicles run freely on the street and get orders from passengers through mobile phone software.