Thousands of high-precision positioning will become an important part of the L3 level of automated driving technology, it can be said that the current market high-precision positioning has become an indispensable part of the launch of high-level automated driving cars by many automobile manufacturers.

These are the first time I've ever seen a car with the same name, and I've never seen a car with the same name.

The world's largest high-precision positioning road test has already achieved milestones.

Recently, the official news of Chance Location said that the road test team has run through nearly half of the country, including Beijing, Shanghai, Wuhan, Hangzhou, Chongqing, Kunming, Chengdu, Changchun, Guangzhou, Xi'an, Lanzhou and other urban expressway, as well as G15, G40, G42, G50 and other highways, with a cumulative mileage of more than 90,000 kilometers.

*Thousandseeker Location's road test in China

This large-scale road test effectively verifies the stability of high-precision positioning services in multiple scenarios, such as different latitudes, different elevations, different road conditions, different weather, etc., and provides massive data support for the optimization of intelligent driving algorithm iterations.

Road test results show that in Hangzhou, where the high-precision map covers, Chance Position can locate which specific lane the vehicle is driving in and provide lane-level navigation information.

*Thousandseeker's road test in Hangzhou

In Guangzhou, the high-precision positioning results of Thousandseeker's location are still accurate despite the serious interference from the ionosphere, which leads to large errors in satellite positioning.

*Thousandsearch Positioning road test in Guangzhou

In Chongqing, even though there are a lot of side roads, elevated roads, and tunnels, Thousandsearch Positioning can still provide accurate, safe, and reliable positioning results. Especially in the case of a long time of star loss in tunnels, Thousandsearch Positioning's self-developed combination of satellite + inertial guidance algorithms still ensures high-precision positioning results, which makes full-scenario automated driving possible.

*Thousandseeker's road test in Chongqing

Nian Jinfei, General Manager of Intelligent Driving Division of Thousandseeker, said, "Large-scale road test is even more long-term and continuous, and through large-scale road test, high-precision positioning can provide a very high accuracy of the location of the information output, which is just needed for the L3 level of autonomous driving, and plays an additional role in L2 level of assisted driving. L2 level assisted driving plays an added role."

How did Chance Position optimize its positioning algorithm to generally improve automotive positioning accuracy to 20 centimeters?

Link: "Algorithm, positioning the more you use the more accurate the secret, here?

Link: "Algorithm, the secret of the more accurate positioning, here?

The first large-scale road test of high-precision positioning in the industry was carried out on all highways and major urban highways across the country to verify the algorithms, which signifies that the high-precision positioning mass-production technology will be landed on a large scale in the field of intelligent driving to provide the security of space-time intelligent infrastructure for intelligent driving.

Thousandsearch's large-scale road test is more than halfway through this period, and the test vehicles, equipped with Thousandsearch's self-developed spatial and temporal intelligence algorithms and high-precision positioning services, have been running all over the country's highways and major urban highways, testing in multiple scenarios for different weather and altitudes, and continually verifying and iterating the algorithms and services.

Previously, Chance Location has completed 300,000 kilometers of high-precision positioning road tests.

According to the current test results, the real-time dynamic positioning accuracy of the test vehicles can reach up to 2cm, which can well meet the wide range of needs of intelligent driving and vehicle-road coordination above L3 level, and realize lane-level navigation. The so-called lane-level navigation refers to the fact that the vehicles not only know which road they are driving on, but also are able to distinguish which lane they are driving on.

So how does Chance Location realize high-precision positioning for large-scale road testing?

First of all, in terms of infrastructure, based on the BeiDou satellite navigation system, Chance Location has built 2,800 ground-based augmentation stations across the country, forming a "national network" of ground-based augmentation systems.

In July this year, Chance Position has announced that "one nationwide network" has been upgraded again, and on the basis of the previous coverage of most areas of the country, it has realized the simultaneous full coverage of the national highway network, including the G7 highway connecting Beijing and Xinjiang, which has just been opened, to "pave the way" for intelligent driving. "

For the first time in the history of the world, the world's largest highway network has been opened to cover the entire country.

For the highway network under planning or construction, Chance Location will continue to follow the pace of national road network construction and synchronize the coverage of Beidou ground-based augmentation stations.

It is worth mentioning that the redundancy and backup between stations have been fully considered in the design and site selection, so that when the station has problems, the nearby stations can carry out redundancy and backup to ensure that the neighboring service capacity is not degraded.

At the same time, the dual antennas and dual receivers are used in the construction of the station, and the full redundancy transmission link is guaranteed in the link of each base station transmitting data back to the data center.

The next step is to let the test vehicles equipped with Chance Position's professional services and intelligent algorithms run on all the highways and urban freeways in China, so as to derive data from the actual large-scale tests, and then iteratively upgrade the algorithms.

*Thousandsearch Location technicians install road test equipment

Thousandsearch Location technicians revealed, "The camera and radar in the vehicle may not recognize accurately in the case of bad weather or lane line defacement, which will lead to a certain safety hazard, and at this time, if it can be combined with the absolute positioning, i.e., satellite positioning, the traveling vehicles will not be afraid of being misled, in professional terms, the relative positioning and satellite positioning combined together."

After the road test data collection is completed, Chance Location will first store the massive road test data in the cloud, and then the server in the cloud carries out automated categorization of the data: including four categories: satellite positioning data, differential correction data, vehicle dynamics sensing data, and high-precision benchmark data.

After storage and classification, the data will be transferred to Chance Position's algorithm library, which is equivalent to a 7*24-hour big data testing ground with many complex mathematical models, matching all kinds of road test data in the algorithm library and performing massive arithmetic analysis to constantly adjust and eventually solve the best mathematical model for high-precision positioning. Through continuous road testing, the algorithm is then tested and iterated, allowing for higher and higher positioning accuracy.

Increasing the database of typical scenarios, high-precision positioning without dead ends

Taking the actual test in Hangzhou Yunqi Town as an example, in the video first announced by Chance Position to the public, we can see that the location of the vehicle driving can be clearly seen through the high-precision map positioning capability of Chance Position, and the technicians demonstrated the lane-level high-precision positioning capability through the tablet PC.

*Thousandsearch Location's high-precision positioning track comparison based on high-precision maps in Hangzhou

In the demonstration, the blue line is the track of high-precision positioning, and the red line is the track of normal positioning, in which the blue line is aligned with the real-time position of the vehicle, and when the vehicle changes lanes, the blue line will change lanes along with it, and the red line has already drifted out of the road.

*Satellite signal-to-noise ratio monitored by Thousandseeker in the urban canyon scenario test

Meanwhile, in the test vehicle, you can observe the signal-to-noise ratio of the current visible satellite signal in real time through Thousandseeker's self-developed software installed in the cell phone, when the satellite signal is lost or blocked, which leads to an increase in positioning error, Thousandseeker solved this problem well through the use of a combination of tight, multi-constellation multi-frequency bands as well as multi-sensor Information fusion is a good solution to this problem, and, improve the availability and reliability of positioning.

Thousand Positions uses tight combination to accelerate the convergence speed, thus realizing the comprehensive improvement of fix rate, convergence time, accuracy and completeness.

*Thousandsearch position in Guangzhou high-precision positioning track comparison

And in different latitudes, different climatic environments, such as Guangzhou, near the equator of the low-latitude region of the ionosphere is very active, it will have an impact on the satellite positioning, resulting in inaccurate positioning and other issues, and the Thousandsearch position test vehicle equipped with optimized algorithms can achieve accurate positioning.

In addition, Chance Position has established a database of dozens of typical scenarios including tunnels, ramps, toll stations, elevated short cover, etc., with more than a million kilometers of mileage, which is combined with the terminal's scenario-aware algorithms, to help the vehicle further improve its positioning accuracy.

*Thousandsearch position in Chongqing Huangjuewan Interchange road test live

Tunnel is a difficult point of satellite positioning, because the satellite signal is blocked, the satellite will be out of lock for a long time, Thousandsearch position using a self-developed satellite combination of inertial guidance, the program is specifically for the development of the optimization algorithms for complex terrain.

*Thousands of location in Chongqing Huayan Tunnel road test

Simply put, the combination of inertial guidance is quite for the fusion of LiDAR and the camera solution, is a relative positioning technology, to assist satellite positioning.

When there is a situation where the satellite is out of lock, the program will immediately call the gyroscope in the car, and through the collected speed and steering data, calculate the vehicle's position in the tunnel, solve the problem of blind spots in the tunnel, and in the case of insufficient available stars, it can still guarantee the integrity of the signal and the safety of the function.

It can be seen that after the completion of the large-scale road test of high-precision positioning of Chance Location, it can empower self-driving cars in many aspects, such as improving the stability and reliability of the perception system.

Currently, the mainstream self-driving sensors, such as camera, millimeter wave radar, LiDAR, although also have relative positioning ability, but because they have their own defects in the process of practical application, the biggest hidden danger is the lack of comprehensive information for the self-driving car in advance of the construction of the scene ahead of the ability to rely solely on the accurate sensors to perceive the powerful computation to deal with the problems of the scene, but the lack of complex scenes in the scene, and the lack of the ability of the car in the complex scene. The most important hidden problem is the lack of the ability to build comprehensive information about the scene ahead for the self-driving car, relying solely on accurate sensor perception and powerful computation to deal with the problems encountered in each scene, but lacking the framework to rely on to deal with high-abstraction problems in complex and extreme scenes.

For example, without the assistance of high-precision positioning, this sensor scheme will affect the vehicle's ability to enter/exit the adjacent highway intersection earlier or later, or when exiting the highway intersection, the vehicle will sway due to the lack of accurate positioning. This leads to a significant increase in vehicle accidents and can also affect vehicle comfort.

The performance boundaries of the sensors result in a scene where some of the information is unperceived, and the lack of curvature and slope angles in the road can have a large impact on the longitudinal and lateral planning and control of the vehicle, and the passenger will feel that the vehicle is not as smoothly controlled as the driver is driving it.

In comparison, if the fusion of thousands of location positioning technology can be well avoided by external factors interference, even in bad weather, non-line-of-sight scenarios and other on-board sensors under the coverage of weak satellite signals, still be able to provide accurate, stable and reliable positioning information for the self-driving car, to help the vehicle to determine the current location, so as to provide an overall perception of the current environment, which is convenient for vehicle

IMT-2020?(5G)? The Cellular Vehicle-to-Car (C-V2X) Working Group of the Promotion Group pointed out in its White Paper on High-Precision Positioning of Vehicles that high-precision positioning plays a decisive role in automated driving in bad weather, repetitive scenarios, non-line-of-sight scenarios, and instability of onboard sensors.

Nian Jinfei said: "GNSS-based high-precision absolute positioning is one of the most core basic capabilities of intelligent vehicles, because intelligent vehicles can only know where it is, especially the precise location, in order to make the correct judgment, and then control the vehicle's next behavior. At the same time, high-precision positioning can also play a role in safety redundancy, although the camera, radar and other sensors also have relative positioning capabilities, but they will be affected by the weather, terrain, while high-precision positioning can help the car to achieve accurate positioning in the whole scene, providing a reliable basis for intelligent decision-making."

Not only that, through the fusion application of high-precision positioning and camera, high-precision radar, etc., it can also realize a good sense of redundancy, for example, GAC New Energy EAN V is equipped with high-precision positioning services, relying on the high-precision spatial and temporal intelligent services provided by Chance Position, GAC New Energy EAN V's positioning accuracy reaches the centimeter level, which makes the vehicle encountered in the process of driving the lane line irregularities, lane line When the vehicle encounters irregular lane lines, short-term coverage, roads without obvious markers, and curves with excessive curvature, and the relative positioning solution fails, it can still continue to provide environmental awareness support for the vehicle.

In addition to this, with the world's largest high-precision spatial and temporal service network of Qianxun Location, Xiaopeng Auto's ultra-long-range intelligent coupe P7 has a strong spatial and temporal awareness capability, and the global positioning accuracy of the whole vehicle can reach centimeter level, so that NGP high-speed autonomous navigation driving can be realized on the highway covered by high-precision maps.

A few days ago, the General Office of the State Council issued the "New Energy Vehicle Industry Development Plan (2021-2035)", which explicitly proposes to promote the application of satellite navigation systems such as Beidou in the field of high-precision positioning. The Plan mentioned in the implementation of intelligent network technology innovation project content, to break through high-precision maps and positioning and other core technologies and products. With the continuous improvement of the level of development of automatic driving, more and more manufacturers also recognize that, in order to continuously strengthen the vehicle to cope with the complexity of the road, the ability to deal with large amounts of data, we can not rely solely on the "sensor + control system" solution.

At this point in time, the first phase of large-scale road test of Chance Position has just ended. At this time, Chance Position's high-precision positioning will become an important part of the L3 level of autonomous driving technology, and it can be said that the market's high-precision positioning has already become an indispensable part of the launch of many automobile manufacturers to launch high-level autonomous driving vehicles.

It can be seen that high-precision positioning and body sensor depth fusion will play an increasingly important role, if the evolution of the key factors of intelligent transportation for a control, we can find centimeter-level positioning is to achieve the key to autonomous driving.

This article comes from the authors of the automotive home car family number, does not represent the views of the automotive home position.