- What does “autonomous” mean?
Basically, “autonomous” is the ability of a machine to perform a task without human control. Therefore, “autonomous system” is a machine, hardware or software that performs a number of tasks on its own when activated or operates itself. Generally, it is capable of sensing and reacting to the surrounding environment.
Autonomous systems are not only restricted to unmanned equipment such as robots but are also integrated into many man-controlled systems such as cars, aircraft, etc. Many cars are equipped with the Anti-lock brake, friction and balance control system, electric steering wheel, emergency seat belt and air cushion. The more advanced vehicles can also be equipped with smart driving systems, automatic lane keeping, collision avoidance and self-parking. Modern commercial aircraft have high levels of automation in each stage of the flight.
- Autonomous vehicles
For the development of unmanned vehicles, software plays the most important role. It is like a brain, ensuring the operation of the car. If in the past, cars characterized by engines, gearboxes, actuators, control steering, gasoline… today, it is like a computer. Software and electricity have replaced the functions of mechanical, human and fuel elements. A modern car is driven by 80 to 100 embedded systems, 90% of the car’s innovations are in software, and 100% of the cars will be connected to Cloud.
Autonomous vehicles are no longer an inviolable realm of Ford, Daimler, Toyota, Nissan… but slowly became the realm of software companies. The research on in-vehicle technology focuses on two main areas: lane detection and object identification.
- Lane detection: This issue has been studied for decades. Most lane detection systems have been developed and applied in many luxury vehicles.
- Object identification: This is an important component of autonomous vehicle system. Recently, this technology has made great steps forward as it recognizes static objects such as bicycles, pedestrians, cars, traffic signs… and is coming towards identifying moving objects such as cars/people on the real road…
- Levels of autonomous vehicle
- Level 0: Level 0 depends entirely on the driver behind the wheel to control the features of the car, completely without self-driving. Most of the vehicles on the market today can count into this category, including cars with collision warning feature or blind spot sensors.
- Level 1: There are some features that assist the driver, for example, to control the speed of the car or the direction of the driver, but can not perform both functions at the same time. Example of level 1 is the cruise control feature, which automatically stops the gas pedal when it is set up from the driver. Or Subaru car also features active lane control on the common car. Cars with auto braking feature, are also included in level 1.
- Level 2: Supports drivers more, but autonomous system at this level can not drive the car at all times, including joining lane on the expressway. Some high-end automakers now offer self-controlling of direction and speed at the same time without the need to interact with the driver for a specified period of time (less than a minute). Volvo, Mercedes-Benz, and BMW all provide level 2 functionality, but all require the driver to track the surroundings when the car is moving. Tesla’s Autopilot can also be considered a second-level autonomous feature, which adjusts the steering wheel’s rotation based on the driver’s ability to focus.
- Level 3: Self-driving depending on conditions. Unlike level 2, autonomous vehicles of level 3 can run independently in all situations, but the vehicle will also track the road conditions themselves, but the car will be driven by humans when they can not continue to handle the situation. By definition of SAE, level 3 vehicles will require people to intervene when the self-driving system fails, but according to some car manufacturers, they think that this could affect the safety of the driver, some manufacturers such as Ford or Volvo announced that they will skip this stage.
- Level 4: Nearly self-driving. At level 4, the car will no longer need the driver’s interaction and the system will stop automatically when the fault is detected. Some automakers, such as Tesla or Mercedes-Benz, have built-in automatic deceleration or light warning if the driver is distracted. Recently, Ford and Volvo have announced that they will launch level 4 vehicles before 2021.
- Level 5: Self-drive. This is the final level and if the car reaches this level, it can drive in all conditions, without the presence of a driver. Levels 4 and 5 are a leap and require a lot of technological innovation. Level 5 does not need humans, it handles itself when a situation occurs. Basic components of the car such as accelerator, steering wheel, the brake pedal will not be needed on the self-driving car of this last level, as there is no person on it anymore. No car maker has given a specific timetable for this last level but only set a general timeline of about a decade.
Increasing levels of automation or self-driving capability can bring many benefits, such as increasing safety and reliability, improving response time and capacity, reducing the human workload to save on expenditures, and the ability to conduct campaigns when communications are interrupted or in harsh environments.
- Opportunity from autonomous car
According to MC Kinsey, like phones, cars will also become increasingly intelligent. Cars will be equipped with driver assistance features such as warning alerts, information on the road, entertainment features,… Mc Kinsey predicts that by 2020, these vehicles will account for 22% of the world’s car, compared with 3% in 2011 and 12% in 2016. The annually average growth rate of this vehicle reaches 30% compared to 4% in tradition.
According to Market & Research, the market of the car equipped with Advanced Driver Assistance System (ADAS) in 2016 has a revenue of $12 billion by 2025 with revenues of $62 billion, a medium growth of 18% a year.
There are more than 30 companies investing in research and development of driverless vehicles such as Google, Tesla, Apple, Audi, Nissan, Nvidia… They have been investing heavily in this area:
- Toyota has invested $ 1.2 billion in developing artificial intelligence in their car lineup with the goal of bringing the car to life in 2020.
- Japanese automakers are also working together to implement 3D mapping for driverless vehicles…
- Audi has spent $3.1 billion to buy Nokia’s street data base.
- GM has bought Sidecar’s assets and invested $500 million in Lyft.
- Uber China says that they will invest 500 million dollars in the idea of making a global map.
- Autonomous technology in FPT
Currently, FPT is implementing a number of projects related to autonomous car technology for major customers in the automotive sector. In these projects, FPT has cooperated with customers to research and develop the features of autonomous vehicles such as:
(1) Video captures hundred thousand kilometers of highways around the world and is tagged display components. This is a very important stage in the development of artificial intelligence, supporting learning machine, especially deep learning.
(2) Developing communication devices (called gateways) of in the vehicle, which will connect the car to the cloud, ensuring that it is always connected to the Internet via 4G / LTE.
(3) Bringing the information collected from the self-driving system (from cameras and sensors) to the cloud for processing centers.
(4) Users can contact the processing center in case if emergency and then rely on the information uploaded on the central server provides a processing direction or the center will automatically propose to the users in case of detecting problems or understanding customers. For example, the center will recommend and send a list of the nearest repair station to the car if it has a problem…
In addition, FPT also participates in the project of communication technology between cars and smart devices through technologies such as V2X, V2Home… and continue. investing in researching autonomous vehicles.
Bich HaiRelated posts: