Technology related to "cognition"

In autonomous driving, it is first required to know the situation around the vehicle. It is to accurately grasp the position where the vehicle is traveling by detecting surrounding vehicles, pedestrians, traffic signs, etc. For that purpose, techniques such as object recognition and position identification are used.

The sensor, which plays the role of the "machine eye" in object recognition, is the most important element. Sensors with different characteristics, such as cameras, LiDAR, millimeter-wave radar, and ultrasonic sensors, are used together to accurately detect the surroundings of the vehicle and hundreds of meters ahead in the direction of travel.

By combining the camera, which is good at identifying objects, with LiDAR, which is good at measuring the distance to objects, the presence of surrounding vehicles and pedestrians can be recognized, and collisions can be prevented by grasping the distance. In addition to preventing traffic jams, it can also read white lines and signs on the road to help control vehicles.

​AI that has learned the characteristics of various objects is responsible for recognizing what the target object is.

GPS and other satellite positioning systems (GNSS) are the main technology used to identify the vehicle's position. In recent years, the quasi-zenith satellite system “MICHIBIKI” has been put into practical use, and positioning with an error of several centimeters is now possible. In addition, technology that uses LiDAR to estimate the vehicle position from surrounding feature points is often used together.

In preparation for tunnels and other environments where satellite positioning systems are difficult to function, an odometer (DMI) that measures distance from tire rotation and an inertial measurement unit (IMU) that measures angular velocity and acceleration are also used.

In recent years, SLAM technology has also attracted attention. This technology performs self-position estimation and map creation at the same time. Based on the relative positional relationship with a landmark, it calculates the amount of movement of the vehicle while mapping the surroundings.

Furthermore, sensors and AI are required to have technology that recognizes objects and predicts their movements. For example, technology that quickly predicts the acceleration and deceleration of the vehicle in front based on information such as brake lights and traffic signals ahead. Predicting the behavior of surrounding vehicles, bicycles, and pedestrians and using it as a basis for decision-making can improve the safety of autonomous driving.

In addition, it will be possible to dramatically improve the accuracy of the autonomous driving system by combining the vehicle's position information and sensor data with highly accurate 3D map information including road signs and lane information. .

Technology related to "judgment"

Next, it is necessary to determine the next action based on various information recognized by the autonomous driving car. AI technology is used as the role of the "brain" in autonomous driving to determine routing, lane selection, and even acceleration and deceleration.

In general, first, based on the information to the destination, an action plan is created, such as which route to travel. It is the same image as a car navigation system. After that, while actually driving the route, changes such as route changes and lane changes are made as appropriate based on traffic information.

In particular, when changing lanes, the system predicts and understands the behavior of surrounding vehicles, creates a candidate trajectory for the vehicle, evaluates each option after considering risks, and determines the optimal driving route. do.

In addition, based on the movements of surrounding vehicles and pedestrians and prediction technology, the robot will be able to ``reduce the speed just in case'' and ``drive slightly to the right (to the left) of the lane if possible'', just like humans. Ability to make decisions flexibly according to the occasional situation is desired.

Technology related to "control"

Finally, it is necessary to control the steering, accelerator pedal, etc. according to the judgment made by AI based on cognition. AI's decisions are converted into electrical signals, sent to actuators that convert them into mechanical movements, and connected to vehicle control. At this time, by-wire technology that replaces the mechanical operation of steering and accelerator pedals with electrical signal control will also become essential.

In addition, remote control technology has also been actively developed. A system that not only monitors but also controls the vehicle from a remote location will be very useful when the autonomous autonomous driving becomes difficult to control for some reason. Remote control requires a low-delay and stable communication technology that monitors images from a vehicle-mounted camera in real time and immediately transmits control commands.

Levels 1 and 2 are equivalent to ADAS (Advanced Driver Assistance Systems)

Level 0 is 'no driving automation', where the driver performs all dynamic driving tasks. Older cars are at this level.

Level 1 is "driving assistance", in which the system continuously performs either longitudinal or lateral vehicle motion control subtasks of the dynamic driving task in a specific limited area. This is the case with either longitudinal driver assistance systems such as Collision Mitigation Braking or Adaptive Cruise Control, or lateral driver assistance systems such as Lane Keep Assist.

Level 2 is achieved by having both vertical and horizontal orientations. Level 2 is "partial driving automation", in which the system continuously performs both longitudinal and lateral vehicle motion control subtasks of the dynamic driving task in a specific limited area.

When level 2 is advanced, "hands-off" driving, which allows you to take your hands off the steering wheel under certain conditions, becomes possible. Since it is a driving support system, the driver must not neglect to monitor the surroundings of the vehicle, but if such technology evolves further and reaches a technological level that does not require the driver's obligation to monitor, it will be "autonomous driving". .

Level 3, the first step towards autonomous driving

Level 3 is "conditional driving automation", where the system continuously performs all dynamic driving tasks in a limited area. While the system is operating, the driver is able to "eyes off" driving, which exempts them from the obligation to monitor the surroundings of the vehicle, allowing them to relax to some extent inside the vehicle. Sleeping and drinking alcohol are strictly prohibited because you have to carry it.

In Japan, the ban was lifted in April 2020 due to the revision of the Road Traffic Law and the Road Transport Vehicle Law, and international standards were established in June of the same year. In addition, in April​ ​2023, the ban on autonomous driving level 4 was lifted in Japan.​

Currently, international standards limit the operation of autonomous driving systems to driving on motorways such as highways at speeds of 60 km/h or less. ODD/operation design area) is also expected to expand.

Level 4 allows driving without a driver

Level 4 is "advanced driving automation", in which the system continuously performs all dynamic driving tasks and responses to difficult-to-continue operations in a limited area. ODD conditions are set for each autonomous driving system, and autonomous driving without a driver is realized. After level 4, the driver no longer needs to think about driving behavior, so it is often called "brain off".

Practical demonstrations of level 4 autonomous driving transportation services are being actively conducted in various parts of the world. Most of them operate with a safety driver on board just in case, but in Arizona, Google-affiliated Waymo (Waymo) has started providing de facto Level 4 service without a safety driver from 2019.

Practical demonstrations are also active in China, and a level 4 taxi service will startin August2022. In Japan, the goal is to realize level 4 unmanned autonomous driving transportation services for limited areas in more than 50 locations around 2025, and in more than 100 locations in 2027. As of July​ ​2023, it is only in Eiheiji Town, Fukui Prefecture, but there are places where it is realized in a short section. Regarding private cars and logistics trucks, the goal is to achieve Level 4 only on expressways after 2025.

Level 5 allows autonomous driving under all conditions

Level 5 is "complete driving automation", a technology that allows the system to automatically drive under all conditions, such as driving area, speed, and weather. Even at level 4, the autonomous driving is almost unnecessary, but it will still be operated only under specific conditions. These specific conditions are the main conditions such as road conditions, geographical conditions, environmental conditions, etc. Other conditions such as the presence of personnel for security and regulations on continuous operation time. There is also

However, level 5 is the level at which operation is possible without any of these specific conditions. It can handle all the rain, wind, and road conditions that occur on a daily basis. There are many movements aiming to achieve Level 5 in the 2030s, but the hurdles are extremely high.

The Japanese government is planning a demonstration experiment for Level 5 autonomous driving at the International Horticultural Exposition scheduled to be held in 2027, and has clarified its policy to consider the feasibility of its realization.

benefits to people

autonomous driving technology frees the driver from driving. The driver effectively becomes a 'passenger' and can use the time freely while driving. In addition, by eliminating the concept of "driver's seat", the degree of freedom in the layout of the interior space is increased. As a result, users will be able to enjoy entertainment such as smartphone operation, reading, and watching movies by arranging the display while traveling.

autonomous driving will also significantly reduce the labor costs for drivers of taxis, buses, etc., which will lead to lower costs for passengers in various transportation services. According to US research firm Ark Investment, the cost borne by passengers will drop to one-tenth if taxis become autonomous driving.

If autonomous driving become widespread, there will be no need for “driving operations” for transportation, so there will be no need to hold a driver’s license. We will create a transportation society in which everyone, including the elderly who have lost their ability to drive, people with disabilities who cannot drive, and children, can move freely.

Total cost of vehicles by service (Source: ARK Invest)

Furthermore, autonomous driving is not limited to the movement of people, but also affects the movement of goods. In the logistics industry, by making it possible to use delivery services at low prices, there is a possibility that everyday shopping such as nearby supermarkets will be changed to home delivery, making it more convenient.

Benefits to society

autonomous driving cars will also greatly contribute to the realization of a safe road traffic society. In 2022, the number of accidents resulting in autonomous driving or death in Japan will exceed 300,000​. are attributed to driver violations. Since autonomous driving cars, in principle, drive safely in compliance with the Road Traffic Law, we can expect a significant reduction in traffic accidents caused by human error.

In addition, the commercialization of autonomous driving will gradually lead to a decrease in the number of private cars. By realizing efficient movement by sharing various transportation services, the total number of vehicles running on the road will decrease, leading to the elimination of traffic congestion.

In local public transportation, where business continuity is difficult due to chronic deficits, we will provide flexible services such as autonomous driving buses and on-demand transportation that utilizes AI (artificial intelligence) to stabilize the means of transportation for local residents. can be secured by

Furthermore, by applying autonomous driving technology to so-called "working vehicles" such as construction machinery, agricultural machinery, police vehicles, and inspection vehicles, it will be possible to save labor in many ways.

The current state of autonomous driving in Japan is Level 3

As of August 2023, autonomous driving in Japan up to level 3 has been put into practical use as commercial vehicles.

Level 3 in private cars was announced by Audi in Germany in 2017 with the level 3 autonomous driving system "Audi AI Traffic Jam Pilot", which can be installed in the flagship model "Audi A8". 

In 2020, when the international standards for Level 3 were established, the movements of automobile manufacturers became active. In Japan, Honda received the world's first approval in November​ ​2020 for the new model "Legend" equipped with level 3 autonomous driving function, followed by Mercedes-Benz S class and EQ.

Regarding autonomous driving buses, France's Navya sells small, low-speed EV autonomous driving buses, which are being used in practical demonstrations in Japan and other parts of the world. In Japan, autonomous driving autonomous driving that travel on fixed routes are expected to be put into practical use earlier than self the Company driving taxis, and from November​ ​2020, in cooperation with Macnica and BOLDLY, we will begin operating transportation services on public roads in Sakai-cho, Ibaraki Prefecture. Legally it falls under Level 2, but technically it is practically equivalent to Level 3.

Level 4 is targeted for commercialization in Japan in 2025

In the development and commercialization of Level 4, Chinese companies such as Waymo, the leading company in autonomous driving taxis, Internet giant Baidu, and startups WeRide, Pony.ai, and AutoX, are advancing technology very quickly.

In June 2022, Baidu will start accepting reservations for an SUV prototype model equipped with Level 4 autonomous driving in China, and the world's first level 4 autonomous driving vehicle will be released. Also, in August​ ​2022, a completely unmanned taxi service will appear, which will also lead the world.

In addition, in April​ ​2023, the ban on autonomous driving level 4 was lifted in Japan. It has begun to operate in limited areas and areas such as patrol buses.

According to the "Public-Private ITS Concept / Roadmap" formulated by the government, in mixed spaces such as community roads, and in limited spaces such as abandoned railway tracks and BRT (Bus Rapid Transit System) dedicated sections, safety drivers will operate at about one location each. It will start a service with monitoring and gradually expand the target, and has set a goal of starting an autonomous driving service with only remote monitoring at several locations between 2023 and 2025.

For general passenger cars, based on the realization of Level 3, we expect to commercialize Level 4, which enables autonomous driving from the entrance to the exit of expressways, by 2025. In order to realize this, in 2024, we are planning to install lanes dedicated to self autonomous driving cars on the expressway (Shin-Tomei Expressway Surugawan Numazu Service Area- Hamamatsu Service Area) in the late-night hours as a support measure.

The logistics service responsible for delivery is also planned to be realized after 2025 in a limited area by applying the technology of autonomous driving transportation service.

(Source:IT Strategic Headquarters "Public-Private ITSInitiative/Roadmap")

Demonstration experiments and examples of buses

autonomous driving of buses has already been demonstrated by many companies under limited conditions on specific routes and areas, and has been released to the public.

Macnica is also conducting demonstration experiments on autonomous driving buses and efforts to utilize Maas in the health and medical field. In Japan, in 2020, the first autonomous driving bus in Japan began regular operation in Sakaimachi, Ibaraki Prefecture.

Also, in 2021, in collaboration with Mitsubishi Corporation and Konan Health Innovation Park, etc., the state of automatic driving of the shuttle bus to the hospital will be released to the media. In this initiative, we experimented not only with autonomous driving autonomous driving, but also with healthcare MAS, in which the patient's health condition was measured and interviewed at the time of pick-up and drop-off, and the results were sent to the hospital in advance.

In February 2023, in collaboration with Mitsubishi Corporation overseas in Indonesia, we are working to experimentally verify autonomous driving buses in shopping malls.

Demonstration experiments and examples of taxis

The taxi industry also has a sense of crisis about the current situation where the operating rate is declining due to the shortage of drivers, and is recommending efforts toward the use of autonomous driving.

In addition, on the technical side, each business operator is proceeding with verification of autonomous driving taxis using the ``Heijo Palace Ruins Historical Park Smart Challenge'' that the Ministry of Land, Infrastructure, Transport and Tourism has been implementing since 2019. In the autonomous driving taxi demonstration experiment, our company Macnica also participated in the operation planning and operational service verification.

On the other hand, in August​ ​2022, Beijing, the capital of China, has reached the level of operating fully unmanned autonomous driving with passengers. 

Demonstration experiments and examples of trucks

The truck industry, like the taxi industry, has a sense of crisis about the shortage of drivers, and the government and others are recommending efforts toward demonstration experiments.

So far, the platooning technology has been successfully tested on the section between Hamamatsu Service Area and Enshu Morimachi Service Area on the Shin-Tomei Expressway. In this experiment, only the first truck has a driver, and the following two trucks follow the first truck without a driver. Therefore, it has not yet reached level 4 autonomous driving.

However, in fiscal 2024, a lane dedicated to autonomous driving will be set up on a section of the Shin-Tomei Expressway (between Suruga Bay Numazu SA and Hamamatsu SA), and in fiscal 2025, a demonstration experiment at autonomous driving level 4 is planned. .

Expressway truck demonstration tests are only conducted at night when there are few vehicles. The goal is to put it into practical use after 2026.

Demonstration experiments and examples of light vehicles

autonomous driving is also progressing in light cars, which can be said to be the means of transportation of the people.

In March 2023, some mini vehicle manufacturers reportedly conducted Level 3 demonstration tests on general roads instead of highways. There is also a background that mini vehicles are rushing to put them to practical use on general roads for residents in rural areas where public transportation is scarce and roads are narrow.

Rise of startups and venture companies

The rise of startups is also remarkable. Starting with companies such as WeRide that develop autonomous driving systems, companies specializing in the development of elemental technologies such as LiDAR development and AI recognition technology are also emerging one after another.

In the autonomous driving field, where highly specialized cutting-edge technologies are gathered, there is a strong tendency for companies in various fields to form partnerships to build autonomous driving cars, and the structure is changing from vertical integration to horizontal division of labor. Impression.

On the other hand, automobile manufacturers are also acquiring and investing in startups one after another, and there is also a movement to revert to vertical integration by bringing newly emerging players under their umbrella.

In Japan, startup ventures such as Tier IV, which leads the development of autonomous driving software "Autoware" and Preferred Networks, which is involved in AI development, are expanding their fields of activity.

CASE and MaaS

Currently, the hottest keywords in the automotive industry are “CASE” and “MaaS”. CASE is a coined word that takes the initials of Connected, Autonomous, autonomous driving & Services, and Electric, and serves as a development guideline for the industry.

Connected refers to vehicles equipped with communication functions, and even ordinary privately owned cars can be operated via smartphones, telematics insurance, enhanced entertainment functions, etc.

Sharing services refer to sharing services that utilize automobiles, such as car sharing and ride sharing.In the future, automobiles will change from being “owned” to being “shared” and “used.”

"CASE" announced by Daimler at the motor show (Source: Mercedes-Benz)

MaaS is related to the viewpoint of using this car. It means a service of transportation, and it seamlessly connects each means of transportation such as trains, taxis, buses, single-seat ultra-compact mobility and kickboards that are expected to spread in the future, and one-stop reservation and payment for use. It is an effort to improve the convenience of transit, etc.

It is expected to appear frequently along with autonomous driving as a keyword that indicates the near future of the automobile industry and the transportation industry, so let's keep it in mind.