Lane-level navigation becomes a new beacon for industry development, and mainstream car companies are accelerating mass production

With the rapid development of automobile technology, people's demand for driving safety and comfort is constantly increasing. As a high-frequency application in automobiles, in-vehicle navigation has also begun to show new development characteristics, moving towards higher information refinement, positioning accuracy and more concrete information display. Lane-level navigation is born from this.

The biggest feature of lane-level navigation is that it can achieve lane-level positioning and lane-level guidance, which can provide users with more refined navigation guidance than traditional navigation. Based on this, car companies including Volkswagen, Mercedes-Benz, BMW, Honda, etc. have begun to install lane-level navigation as an important configuration on new cars, and car navigation has entered a new era of advancement to lane-level navigation.

Meixing Technology is the world's first mass-produced lane-level navigation

Meixing Technology is the largest independent navigation software provider in China and an important member of the NDS Association (the NDS Association was established by BMW, Volkswagen, Daimler and other internationally renowned car manufacturers, with the main purpose of providing a set of standard international map data formats for the future development of car manufacturers, system suppliers and map suppliers). Meixing Technology has played an active role in the formulation of the navigation data standard NDS. Its navigation engine and data compilation platform based on the NDS standard have global adaptability and quickly gained the favor of mainstream international car manufacturers.

As the next hot spot for in-vehicle navigation, lane-level navigation has entered the stage of large-scale application under the active promotion of Meixing Technology and others. Meixing Technology's lane-level navigation has been applied to new cars of many mainstream car companies since its mass production at the end of 2019.

Image source: Meixing Technology

What supports the rapid mass production of lane-level navigation by Meixing Technology is its rich R&D and project delivery experience accumulated over many years in the navigation field. Since the official launch of lane-level navigation product development in 2016, Meixing Technology has independently developed a number of key technologies, including lane-level positioning algorithms, MXNAVI lane-level map generation system, lane-level AR, scene rendering, etc., building a solid barrier for the large-scale application of lane-level navigation.

Meixing Technology's mass-produced positioning algorithm can provide lane-level sub-meter positioning results by integrating GNSS information, vehicle body sensor information, multi-function camera signals, etc., which can refine the positioning accuracy of the car from the road level to the lane level. Meixing Technology said that the lane positioning accuracy and lane change recognition rate of its lane-level navigation products currently exceed 95%. Using this positioning solution, even in scenes with weak GPS signals such as elevated roads, underground parking lots, urban canyons, tunnels, etc., it can still provide accurate and coherent spatial positioning. The positioning recognition rate of multi-story elevated bridges reaches 96%, and the accuracy of multi-story underground parking lot floor positioning can reach 100%.

Image source: Meixing Technology

The improvement of positioning accuracy makes the driver more sensitive to the spatial relationship between his own position and the surrounding environment, providing an important decision-making basis for the driver or the autonomous driving system to plan and control the next step. This multi-source fusion positioning solution can also greatly improve positioning redundancy safety and enhance the robustness of the navigation system in complex scenarios.

Based on lane-level positioning and lane-level data, Meixing refines navigation route planning and guidance from the road level to the lane level. Lane-level navigation can provide users with more appropriate and accurate lane guidance. During driving, it clearly informs the driver of lane selection and lane changes in advance to avoid the driver's troubles such as untimely lane changes, detours, or safety risks. For example, in the road design of some cities, the U-turn lane is in the rightmost lane. If there is no guidance from lane-level navigation, it is easy to habitually choose the leftmost lane. This is a typical application scenario of lane-level navigation.

Image source: Meixing Technology

The MXNAVI lane-level map generation system mainly processes standard SD map data into high-precision HD data that can be used at the lane level through software. Through this system, Meixing Technology only needs SD map data and a small amount of HD lane attributes to achieve the fusion of SD maps and HD maps and generate map data that can be used for lane-level navigation. In this way, lane-level navigation does not necessarily have to rely on full-factor HD map data, which lowers the threshold for car manufacturers to evolve from traditional navigation to lane-level navigation. This makes Meixing the world's first software company to make up for the lack of map data through software, and it has also greatly accelerated the large-scale mass production process of lane-level navigation.

Image source: Meixing Technology

Scene rendering makes the lane-level navigation information presentation more intuitive. Meixing uses the SD road network and related lane attributes to identify different models and generate lane 2D vector graphics. Then, by further adding the road network topology and ADAS attribute information, it derives the relative depiction hierarchy of the road and generates 3D modeling. After that, it only needs to combine the connection relationship of the lane graphics and the changes in lane attributes to produce vivid and intuitive auxiliary guidance information. Based on this technology, Meixing Technology can achieve rendering and visualization of lane-level navigation.

Meixing Technology's lane-level navigation is a significant improvement in the experience of traditional in-vehicle navigation. Meixing mass-produced highway/urban express lane-level navigation functions in 2019, and launched urban trunk road lane-level navigation functions in 2020. According to Meixing Technology's plan, in the next five years, the company will complete the generation, compilation and launch of lane-level data for trunk roads in 56 cities across the country to support the application of its lane-level navigation on a larger scale.

The evolution of advanced navigation in the US continues to drive navigation reform

With the continuous evolution of perception products and technologies such as cameras and radars, the capabilities of lane-level navigation are also expanding. At present, based on the existing lane-level navigation, Meixing Technology is also actively conducting technology research and development to promote the integration of navigation engines and vehicle perception information to form a series of high-end navigation products for human-machine co-driving.

1. AR navigation that integrates lane-level positioning is the real AR navigation

Lane-level AR navigation is a fusion of lane-level navigation and visual sensor information, and its essence is still human navigation. AR navigation does not mean that as long as the picture is dazzling, it will be meaningful only if it is based on lane-level positioning. Meixing's lane-level AR navigation integrates high-precision lane-level positioning and image recognition technology to achieve a perfect fit between the trajectory line and the actual road, and integrates route guidance information, ADAS information, etc. into the driver's perspective. Lane-level AR navigation can not only provide WYSIWYG route guidance, but also realize a series of driving assistance functions such as pedestrian reminders, curve reminders, and anti-collision warnings, improving navigation experience and driving safety.

Image source: Meixing Technology

In the next stage, Meixing’s newly developed technology will integrate lane-level AR navigation, image recognition, SLAM positioning, AR cloud anchor technology, and lane-level scene interaction, making AR navigation infinitely close to MR navigation.

2. High-precision lane-level navigation is a window for human-vehicle interaction during the co-driving stage

In recent years, autonomous driving technology has developed rapidly. Whether it is single-vehicle intelligence or vehicle-road collaboration, it will take a long time to achieve fully unmanned driving, and human-vehicle co-driving will still be the main form of mass production. In this case, how to better involve drivers and effectively improve driving safety is an important topic.

Cruise control on highways is probably the earliest known autonomous driving assistance function. In this scenario, there is no need to set a destination, and the car only needs to automatically maintain the speed within the driver's visual range. To achieve autonomous driving functions in route planning scenarios such as automatically selecting the optimal lane, automatically switching on and off ramps, etc., it depends on high-precision lane-level navigation. This involves improvements in high-precision maps, high-precision positioning technology, visual sensors and radar engine algorithms, and HMI interaction.

Image source: Meixing Technology

Lane-level navigation will be the main interface for interaction between drivers and cars during the co-driving stage. With the help of 3D rendering technology, the lane-level navigation interface can not only provide drivers with precise positioning and more intuitive and clear lane-level navigation guidance, but also make the spatial position relationship between the vehicle and the surrounding environment such as vehicles, pedestrians, and obstacles more concrete. In the state of autonomous driving, the lane-level navigation interface can not only display the car's perception of the environment, but also present in a visual form the decision intention and execution action made by the autonomous driving system based on the lane-level navigation route and perception information. The visual presentation of the autonomous driving decision-making process is the basis for the driver to establish trust in the car's autonomous driving system, and it also allows the driver to predict the behavior of the autonomous driving system so that he can quickly take over control of the vehicle in certain emergency situations.