Hands-off is an inevitable trend in the development of the automobile industry. With more and more intelligent driving functions becoming the factory standard of cars, how to improve the safety and accuracy of self driving cars? What will support overall path planning for autonomous vehicles? These have become the consensus questions of automobile industry.
We take the realization of automatic parking in the restricted scene as an example. The high-precision combined navigation system relies on the surroundings data returned by the sensor system. The processing center replaces the human brain to issue commands to the control system. The automatic parking is finally realized through the actuator. However, with the expansion of the parking scene, the road complexity is constantly increasing. The data collected by sensor systems alone is not sufficient to support higher level driving strategies.
Role in Autonomous Driving
If the map navigation and vision sensors are compared to the “eyes and ears” of the car, the high-precision combined navigation system can be regarded as the “cerebellum” of the unmanned driving system, adding the ability of “overall control and balance” to the car. It is usually based on GNSS to obtain positioning satellite signals, supplemented by differential signals from ground reference stations, combined with INS measurements of the vehicle’s own three-axis acceleration and three-axis angular rate for trajectory estimation. These three methods work together to accurately capture information about the vehicle surroundings, road conditions, and travel trajectory.
Why GNSS+RTK+INS can give the car the ability to “overall control”?
GNSS is also known as the global satellite navigation system. It provides users with all-weather three-dimensional coordinates, speed, space & time information, such as China’s BeiDou and the United States’ GPS. High-precision GNSS measurements require the use of carrier phase observations.
RTK/Real-time kinematic positioning
RTK positioning technology is a dynamic positioning technique based on carrier phase observations. It can perform real-time differential corrections, provide three-dimensional positioning results in the Earth coordinate system, and achieve centimeter-level accuracy.The high-precision differential satellite positioning technology is highly dependent on the wireless signals and communication links.
If driving in the city canyons, tree-lined roads, under the overpass and other signal poor scenes, inertial navigation system will play a very important role in driving when the car is “blind”. When there is no GNSS signal, INS mainly relies on the three-axis MEMS acceleration sensor, three-axis MEMS gyroscope sensors and solution circuit in the inertial measurement unit (IMU). The IMU outputs the positioning and attitude data of the vehicle based on the DR Algorithm, and calculates the navigation status at the next moment.
IMU/Inertial Measurement Unit
Calibration technology helps IMU achieve maximum performance
High-precision IMU has a characteristic in the performance requirements, that is, “there is no end”. Besides the automotive industry, IMU is also used in areas such as robot autonomous mobility, such as vision-based real-time positioning and map construction (VSLAM/VIO). The VSLAM/VIO acquires RGB and map deep information based on cameras, but also integrates IMU sensors for positioning assistance.
DAISCH did some performance comparison tests. Two MEMS IMUs with Gyro zero bias instability of 3.5°/h and 6.3°/h are combined with binocular vision for VSLAM construction. The test result shows that the trajectory rendering performance of different samples will be quite different with the increase of walking distance (about 550m per circle) under strict control of other variables.
Based on the comparative data of 50 repeated tests, sample A with zero bias instability of 3.5°/h has significant advantages in accuracy and repeatability. Its error shows better stability after accumulation. In the scene without the assistance of wheel speed odometer, the box chart distribution of about 5 meters can still be achieved.
Why is calibration necessary?
IMU outputs vehicle positioning and attitude data based on DR Algorithm. The accuracy of DR Algorithm depends on the accuracy of gyroscope and MEMS accelerometer in IMU. Due to the cumulative error in the integral calculation of the computing module, the error will increase linearly with the increase of time. In addition, in practical applications, high-frequency vibration can also reduce the reliability and accuracy of IMU hardware in INS components. This is the main reason for the increasing demand for high-precision IMU calibration technology in automotive scene.
Based on high-precision standard equipments, DAISCH performs a large number of data acquisition, matrix calculations, and finally feeds back into the IMU processing chip. The experimental data show that the zero bias error with temperature drift can be reduced to 0.01dps after calibration, which is 1-2 orders of magnitude higher than before calibration. In the full range of the sensor, the scale factor error range can be reduced from 3% to 0.3%, an increase of 1 order of magnitude. The orthogonality error can also be reduced from 1% to 0.01%, an increase of 2 orders of magnitude.
Strict quality control of DAISCH products
DAISCH performs rigorous product tests on all devices. The temperature rapidly rises and falls in tens of seconds to get a comparison chart of temperature cycle curves. This tests the effect of deformation and stress on zero-bias stability under high temperature scene, and calibrate the sensor. The IMU sensors are also continuously inspected to ensure the product stability in lifetime and validity of calibration parameter.
DAISCH Product roadmap
In 2015, China’s vehicle market was still in the introduction period of ADAS. The combined inertial navigation technology was mainly in the UAV industry.
In 2016, DAISCH applied IMU, combined inertial navigation and automotive ECU development technology to the UAV field, launching the first UAV control system in UAV industry.
After 2016, unmanned driving market is becoming more and more hot. Many segments of the unmanned driving industry have become investment hotspots. According to the forecast, inertial navigation industry growth space is broad. It is expected to reach 6.6 billion USD in 2026, the compound annual growth rate of 14.50%.
In 2018, DAISCH launched IFS series, the first high-performance test tool in automotive industry. SAIC Passenger Car, PATAC, XCMG and other enterprises use IFS.
Since then, DAISCH has stepped into the forefront in the field of inertial navigation systems. In 2020, DAISCH’s fully automated P-BOX (GNSS+RTK+INS combined positioning mode) assembly and calibration production line is put into operation.
In 2022, DAISCH launched IFS3000, a new generation of high-performance, fully functional, networked combined inertial navigation system. The IFS3000 has desktop software system. Its update rate is up to 100Hz. The precision reached the level of optical fiber combined inertial navigation system. IFS3000 is suitable for autonomous driving development, ADAS test applications and other fields.
For complex scenes such as mines, DAISCH launched the IFS2100 that can withstand fierce bumps. IFS2100 has GB3836, GB/T 4208 explosion-proof certification, meeting the mine explosion-proof safety standards. What’s more, IFS2100 meets the IP67 dust prevention and waterproof level, and can withstand the ambient working temperature of -40 ~ 85 ° C. In addition, IFS2100’s aviation aluminum enclosure is able to withstand harsh electromagnetic interference. The aluminum enclosure is to ensure heat dissipation and to resist radiation interference in mining areas.
In 2022, DAISCH obtained the A-SPICE L2 level certification of software development in the field of high-precision automotive inertial navigation. DAISCH is the first Chinese company to obtain this certification in the field of automotive high-precision inertial navigation. This marks that DAISCH’s software quality system of related products has reached the international leading level.
Sum up
At present, DAISCH develops flight controllers for large-load UAVs, and provides truth value devices for ADAS and active safety systems, ultra-high precision combined inertial navigation for unmanned mining vehicles, high-precision IMU for terminal delivery vehicles, and also provides P-Box products and supporting services for L2+ passenger car of many head automobile manufacturers.
