Autonomous system overview: What is behind autonomous driving?
Autonomous vehicles are our near future, which autonomous driving systems will we see on the highways first?
France is one of the first European countries to establish a complete regulatory framework for the movement of autonomous vehicles. From September 2022, France will start implementing provisions of the highways and traffic laws to allow the traffic of vehicles with authorized autonomous driving systems, as well as the automation of road transport systems on predefined routes or zones.
The first passenger cars equipped with autonomous driving systems will be cars whose dynamic control will be fully delegated to the automatic driving system – but only under certain conditions such as highway traffic jams. Such vehicles could be approved for use on public roads by the end of 2021.
The first approved system of this type is the Automated Lane Keeping System (ALKS). ALKS controls the lateral and longitudinal movement of the vehicle for longer periods without driver intervention.
How does this system work and what is behind the development of such a complex system?
The RT-RK Automotive company, which is part of the TTTech Auto Group, is working on the development of a software platform called MotionWise™ that is responsible for safe autonomous driving.
What should an ALKS contain?
1. A camera mounted on the front* of the car
*in the windshield or in the front mask
In this case, the camera is a sensor that will further send information to the system. What does this sensor detect? In order for the vehicle to "keep to the lane", the sensor must detect the lanes on the road.
As the cameras are often unreliable - they cannot detect bad or damaged tapes, they perform worse in conditions of reduced visibility - they are often combined with additional lasers or infrared cameras.
2. The processor that will process the information that the sensor (camera) sends to it
Every autonomous vehicle has a processor that analyzes images and isolates lines (lanes on the road). Based on the trajectory of the vehicle, it is calculated which correction of the steering wheel and brakes is needed - whether the vehicle needs to turn sharply or brake, or if only a slight correction is needed, in order to stay in the lane. Since the behavior of the vehicle on the road depends on its dimensions and mass, in order to correctly calculate the trajectory, it is necessary to take that into account.
As the "lane keeping" control systems have still not reached the L3 level - they are not yet fully autonomous, the driver can take control of the vehicle at any time.
What it means? At any moment while you are in traffic on the highway, if it seems to the driver that the vehicle is not making the "right decision" in traffic, he can take control of it.
3. Several teams ensuring that everything is designed to the highest security standards
In order to develop such a system, several different teams are needed to make sure that autonomous driving in this vehicle will take place in a way that is safe for all road users. Those are:
● Legal team - they study the current standardization and legal aspects of autonomous driving
● Requirement engineers – study customer requirements
● System architect – performs feasibility analysis
● Safety engineer - performs a safety analysis
● Test manager – creates a verification plan
And of course, the development team develops algorithms for several types of problems:
● Machine vision – extracting useful information from camera images
● Sensor fusion – proper analysis of detected objects and lines and its alignment with the state of the car (coordinate systems)
● Vehicle model – management of the vehicle itself
Safety first
As we can see, in order for autonomous vehicles to appear on the highways by 2022, the most important thing is vehicle safety.
This is why safety testing is included in the entire process of developing autonomous driving systems, such as automatic "lane keeping" systems.
First, safety goals are determined based on risk analysis, as well as on the basis of their severity and the frequency of certain dangerous situations. Based on this, each algorithm, sensor and actuator are assigned the appropriate importance to meet the safety objectives.
In the further development of the system, additional analyzes and refinements of the initial design are carried out until a satisfactory security system is reached - reducing the probability of occurrence of errors to an acceptable level.
All errors that are noticed during the development of the system, which may lead to some danger, are either removed (errors in the design are corrected) or additional monitoring mechanisms are introduced in order to prevent or detect the error in time.
All new security mechanisms are additionally tested through SW and HW simulations.
All current autonomous systems rely largely on the presence of the driver – meaning that most fail-safe systems occur by notifying the driver, who then takes control of the vehicle.
How can you be part of the development of these systems?
A large number of different teams are working on developing these systems so that we can see the first autonomous vehicles on the highways as early as 2022. What would be your tasks as a junior developer in such a development team?
Some of the tasks include implementing and testing the finished architecture, as well as writing the code for each subsystem and testing its validity. For example, writing a part of the algorithm that determines the angle and slope of the detected line or writing a module that receives images from the camera and verifies the timeliness of the received image.
This is just one example of tasks, but as we can see - there are a large number of different tasks for the development of such a system.
So, if you want to be part of the development of advanced systems for autonomous driving and build a career in the automotive industry, follow the RT-RK company profile and stay up to date with the latest job ads.
