Monday, Jul 15, 2024
Update At 14:00    USD/EUR 0,92  ↑+0.0002        USD/JPY 151,69  ↑+0.174        USD/KRW 1.347,35  ↑+6.1        EUR/JPY 164,16  ↑+0.143        Crude Oil 85,49  ↓-0.76        Asia Dow 3.838,83  ↑+1.8        TSE 1.833,50  ↑+4.5        Japan: Nikkei 225 40.846,59  ↑+448.56        S. Korea: KOSPI 2.756,23  ↓-0.86        China: Shanghai Composite 3.015,74  ↓-15.745        Hong Kong: Hang Seng 16.512,92  ↓-105.4        Singapore: Straits Times 3,27  ↑+0.018        DJIA 22,58  ↓-0.23        Nasdaq Composite 16.315,70  ↓-68.769        S&P 500 5.203,58  ↓-14.61        Russell 2000 2.070,16  ↓-4.0003        Stoxx Euro 50 5.064,18  ↑+19.99        Stoxx Europe 600 511,09  ↑+1.23        Germany: DAX 18.384,35  ↑+123.04        UK: FTSE 100 7.930,96  ↑+13.39        Spain: IBEX 35 10.991,50  ↑+39.3        France: CAC 40 8.184,75  ↑+33.15        

Building high-precision, location-based infrastructure as a foundation for a digital society

Interview - January 18, 2024

Revolutionizing navigation, Dynamic Map Platform leads in high-precision mapping for autonomous driving and beyond.


If we could start with an introduction to your company. What is Dynamic Map Platform's main business, and what are some of the core competencies that set you apart from your competition?

First, let me share our Vision. Our company's vision is "Building a high-precision location-based infrastructure globally as a foundation for a digital society and opening up a new future for autonomous driving and other industries."

We provide technologies not only to the autonomous driving (AD) / Advanced Driver Assistance Systems (ADAS) sector but also to other industries. With the use of high-definition (HD) 3D data, we aim to help other sectors drive innovation. Our company was founded in 2016, and we currently have 209 employees on our books. We have office locations in Japan, the US, Europe, South Korea, and the Middle East. Interestingly, we have 25 companies as our shareholders, most of which are related to the automotive industry, including 10 major Japanese automotive manufacturers. We are also supported by the Japanese government through various funds, and our other shareholders come from a variety of different sectors. Currently, our plans include global expansion through acquisitions.

Our company has two main revenue streams, the first being the auto sector, and the other being non-auto sectors through our 3D digital business. The way we create our data is processed through four steps. The first step involves using satellites to identify the locations of our vehicles, and then we do the measurements of these locations. After that, we do the actual mapping, and finally, the integration of data into the desired form.

Many people ask what the difference is between our 3D map data and car navigation data, and our data does not include points of interest and aid human navigation. Instead, we collect road and physical information and express that in a 3D form for precise vehicle control. You might find yourself asking why people need our technology. The first usage case, in my opinion, comes down to safety, and in bad weather conditions, an ADAS vehicle without our HD map finds it difficult to identify objects. This could prove fatal, and this is where our tailored data sets can come in. The second usage case is to improve comfort. Let's take, for example, a road with a very sharp curve. A car with only sensors and cameras would possibly take that turn in a zigzag pattern to optimize the route. By using our technology, that would be taken much more smoothly than a zigzag. While we are providing these technologies to automakers, this technology can be applied to several different applications in different industries. We are seeing an upside in robotics, mobility, and industrial equipment controls.

Let me explain some more usage cases in a little more detail. Imagine a car is reaching an intersection, and the car is supposed to stop at the line. We can tell the car to stop at the line and look ahead to the signal. We can provide exact information to the car, whereas if we rely purely on sensors, that information could be erroneous. By that, I mean that the sensor is looking for the signals, and it might pick up signals for pedestrians rather than vehicles. This could be disastrous when making a stop-or-go judgment. With our data, we can direct the sensors to the exact signals they are looking for rather than taking a guess. In situations where lines are obscured or not there, our mapping data draws the invisible 3D data line for the vehicle. The vehicle will then follow this virtual line. In this sense, we can build virtual boundaries and prevent errors from occurring, and we refer to this as a Geofence.

Currently, the industry has designated 5 levels of AD/ ADAS, from level 0 to level 5. Between levels 1 and 2, it is possible to achieve ADAS with a combination of only cameras and radars. If you advance even further, it is possible to have hands-off driving safely and comfortably. The market for level 2 plus is expanding, and that is enabled with high precision 3D mapping data. At level 3, car manufacturers are liable for accidents, so currently, we are focusing on level 2 plus, and the belief is that it will bring us more volume market-wise. Not only in Japan but globally, the opinion right now is that level 2 plus should be the current focus. However, our technology can cover beyond level 2 plus, all the way up to level 5. The ADAS market, in general, is expected to grow and expand, especially on level 2 plus.

We are expanding our business globally in Japan, the US, Europe, South Korea, and the Middle East, and I believe that most level 2 plus vehicles are utilizing our data in some form. These are just examples of different companies that adopt our data, and we have been allowed to disclose their applications. Nissan, Honda, Toyota, and GM are all huge carmakers that are currently utilizing our technology within different models.

When we established the company, there was a Japanese scientific investigation project (SIP) in place, initiated by the Japanese government. Our company was founded under this initiative, and that is why the Japanese government is a major shareholder. The goal at the start was to explore human-machine interfaces as well as safety and high-precision mapping technology. Those technologies are all necessities in ADAS operations.

Our data covers 35,000 km in Japan, 550,000 miles in the US, and 120,000 miles in Europe. I think we are the number one company in the sector when it comes to the sheer volume of data we possess. Our data also achieves best-in-class accuracy, as much as single-digit cm among map providers, which is essential for AD/ADAS. I think in terms of our strengths, Dynamic Map has a sort of trident consisting of globally large coverage, high precision, and strong partnerships. Compared to competitors, our data is exceptionally detailed and precise, and I think that is why so many companies choose to partner with us to augment their ADAS systems. This isn’t limited to the domestic market either, and in the US, we have very strong connections with GM as well as other US automakers.

We are applying our technology outside of the auto industry to places like the mobility sector for logistics and transportation, as well as infrastructure management, maintenance, repairs, and disaster prevention. 3D mapping can even be used in the entertainment and advertising industries as well. The top one right now is for infrastructure, and for that, we are working together with the Japanese government to gather common data for autonomous driving. We are also collaborating with other private companies such as Telecom company for things such as mobility and robot controls. In the area of infrastructure management, there is a snow removal project in Nagano prefecture, and our technology can be used to assist in snow removal. We are thinking about expanding it not only nationwide but also overseas.

Green innovation comprises all types of innovations that contribute to the creation of key products, services, or processes to reduce the harm, impact, and deterioration of the environment while at the same time optimizing the use of natural resources. We currently have a green innovation project underway that aims to reduce CO2 emissions. In Japan, trucks account for a large amount of CO2 emission annually, so by using our 3D mapping data, it is possible to identify the optimal route for that transportation, cutting down fuel costs and CO2 emissions. Our data can also provide a favorable place for EV chargers along the route, optimizing deliveries from large vehicles such as trucks.

Collecting and processing such huge amounts of 3D mapping data does take a lot of time and a lot of money. Our "Spatial ID" is HD coordinate-based data that expresses location information in actual space. We voxelize the space and assign Spatial IDs so that the Spatial IDs can be used as keys for searching dynamic information and Common Reference Points. Therefore, certain vehicles such as mobility devices, drones, and heavy goods vehicles (HGV) can share the same location data, so in this sense, it can be effectively controlled. The way we have set up our system is that we can provide specific data to specific IDs. For example, a client may only require weather information, and therefore their ID is set up to only receive that data. It is possible to control the flow of data to specific clients, thus eliminating bloat.

As you might be aware, over 28% of people here in Japan are over the age of 65. By mid-century Japan is looking at a population of less than 100 million people, and if we look at the state of roads here currently, safety is a huge concern with the large number of elderly people. Last year 17% of accidents were caused by this elderly population, up from 15%, and the highest on record ever. How does ADAS in conjunction with your mapping software mitigate these risks? How can you take control and ensure that accidents don’t occur as frequently?

Many elderly people have deteriorated perception and recognition because of their increased age, so the problem occurs when they tend to overlook specific road signage or traffic lights. Unfortunately, roadside information and signals are necessary to drive safely. Using our 3D mapping data, we can block vehicles from entering a pedestrian-only street or we can use the Geofence technology we mentioned earlier to block vehicles from entering areas they shouldn’t. In this sense, we can block a vehicle from veering into the wrong lane, which is especially dangerous on expressways.

There is also an issue when elderly people suddenly get ill when driving, and in this case, we can alert the driver or even stop the car in a safe spot. The car needs to be pulled over at a safe space, and to do that we need to understand the width of the road and if a space is safe or not for stopping a vehicle. With the features highly accurate location data is important, and wrong information sent to a vehicle could be disastrous or even fatal.

One of the big challenges when it comes to mapping large areas is making sure that your data is up to date. When there is a change on the road or there is an accident where a barrier breaks, you have to be able to reflect that instantly within the database. How do you make sure that your data is constantly refreshed and up to date as you cover so many miles across the world?

We gather information to identify changes on the road from road management companies, municipalities, and local governments. To add to this information, we also gather information from sensors in cars, which we refer to as probe data. That data is analyzed and cross-checked against our other data sources. When we detect any changes, we use our Mobile Mapping System (MMS) to update information regularly, which is the same system as we collect initial data. The probe data is currently used only to identify changes, but the future goal is to update this information without the need for MMS. It would be more efficient to update data utilizing the probe data, rather than MMS, thus providing quick updates. The main reason why updates are possible with probe data without MMS is because the initial data using MMS is highly accurate and can be referenced to provide confidence in the accuracy of the updated data required by AD/ADAS.

Currently, you are relying on data networks to transmit this data, so perhaps there’s some latency in the transmission of that data. How do you roll out this data consistently with different networks worldwide with different operating speeds, especially for developing countries in the future?

This is why we need HD Maps. There are 4 layers to the HD Map, with the bottom layer being static, the second being semi-static, the third being semi-dynamic, and the top layer being dynamic. Major changes only happen on the bottom, static layer, and those changes only happen on a few percent of the road around once a year. This might include brand-new roads or construction projects. This layer doesn’t change that frequently; instead, accuracy is very important in that sense. In some cases, our map data is updated through Over The Air technology but also updated periodically when Internet access is available or car owners visit the dealer shop for maintenance. The dynamic layers include information such as bad weather conditions or very strong sunlight. By combining both static and dynamic information, we can make more accurate information.

While the probes are reading bad weather information in real-time, we have the static information of the road, the structures, and the physical elements in the HD Map so we can maintain consistency.

One of the big silver linings of COVID-19 in Japan was the fact that many people have moved out of big cities like Osaka and Tokyo to more regional areas, and this has been something that the government has actively promoted. With people now living in rural areas again there are chances to grow those areas, especially for tourism, but we know there are still concerns about accessing those areas in terms of public transport. How can ADAS and mapping software help these areas become more revitalized and more accessible to the public that still lives in big cities like Tokyo or Osaka?

I think you are right that one of the biggest challenges facing national revitalization is the lack of public transportation. Additionally, you still have a problem with depopulation because of the problems with aging. I think by using our technology, we can overcome insufficient public transportation by introducing new mobility tools for people living in those local areas.

The Japanese government is also aiming in this direction based on their 2025 and 2028 initiatives. They are increasing the number of autonomous driving test locations from 50 cities to 100 cities nationwide, and they are also going to set up special lanes for autonomous driving. These tests are Level 4 and Level 5 autonomous driving, meaning that these tests will be conducted driverless, and to achieve this, we believe that accurate mapping data is needed.

Your company is a Japanese firm that owns terabytes of data about American infrastructure, and America is a country that cares a lot about safety and security. Currently, there are rising tensions between the US and China over critical technologies such as yours. Companies like Huawei have been banned because of their access to sensitive data. When you expand your services to foreign countries like the US and Europe, how do you ensure your data won’t jeopardize the safety of citizens in the eyes of local governments and lawmakers?

We work with global law firms to comply with the laws and regulations of each country. For example, some countries have banned firms from other countries from removing data from the country, so in that case, we have set up our cloud-based systems within the country so that we are careful not to remove the data and bring it overseas.

In the political domain, it is difficult to talk about because the information in this regard is very sensitive. When we acquired Ushr in the US, we applied for the necessary applications with the authorities. Japanese companies operating with this kind of data require approval, which we have been granted. It is important to know too that we are not conducting any lobbying activities but rather just complying with local laws.

We are curious to get your opinions on mobility as a service in the future. Do you believe that bus and light rail service will no longer be needed, and cars themselves will become service vehicles? How do you envision mobility as a service playing out by the year 2050?

We believe that the best transportation system differs depending on the environment, such as a highway or a local street, and an urban area or a rural area. Therefore, by 2050, various systems will be used depending on the purpose, such as autonomous driving mobilities, buses, and light rails. For example, autonomous driving small mobilities have begun to be used for tourists in regional cities in Japan. Autonomous driving systems using HD maps will become standard for mobility that requires driving in specific areas and routes at relatively low speeds. This is because HD maps created in a standard format can be installed at a lower cost as data to control mobilities. For ADAS cars as well as mobility as a service (MaaS), we are going to analyze the data and expand our services going forward. We are targeting not only ADAS systems but by providing location information, we can increase the added value of our services, thus satisfying different and varied customers.

Your firm has tackled some very high-profile projects, such as the development of level 3 autonomous driving for the Honda Legend, which is the world’s first level 3 autonomous driving project. Could you run us through some of the projects that you feel most proud of in terms of overcoming a huge challenge that was in the industry, and you were able to break through the barrier standing in your way?

The Honda Legend project was a huge challenge as we broke through to a completely new level of ADAS.

At the same time, increasing the map coverage and updating it regularly is another challenge that motivates us. Green Innovation is another special key element, and the goal of our company isn’t just to achieve autonomous driving but to also reduce CO2 emissions. I think these cases are all linked to each other, and using our data, we can control vehicles and make them more efficient. Using the data we gather, we can plug that info back into the system again, thus creating this circular system of improvement. This is all possible because of our 3D data.

HD maps are attracting attention in the new digital society where the Japanese government is trying to solve various social issues, and we believe that HD maps play an important role in solving these issues. This is a big opportunity and challenge for us.

Imagine that we come back and have this interview all over again in three years. What goals or dreams would you like to have achieved by the time we come back for that new interview?

As a personal goal, I would like to take this company to a whole new level in terms of revenue, the number of employees, and the area of operations. I would like to continue pushing this company forward, growing from strength to strength. There aren’t many Japanese companies that are super competitive in the global market, so I would like to make Dynamic Map Platform one that is. In the business of data, language doesn’t matter as much, so in that sense, I would like my firm to become a competitive startup that is recognized globally.