Autonomous driving is smarter, and redundant sensing is not “redundant”

The driving culture of the "holy city" Jerusalem is known for its wildness. When the hustle and bustle of the city during the day gradually fades into the dim lights at night, the roads of Jerusalem welcome a confident and decisive self-driving car. This driverless vehicle autonomously completes the entire process of online taxi service, not only driving to multiple destinations, but also automatically stopping where passengers are needed.

Recently, Intel's subsidiary Mobileye released its latest autonomous driving video, in which this "smart" car equipped with a "truly redundant sensing system" is the protagonist. It is precisely because of this system that this autonomous driving car can successfully complete the online car-hailing shuttle service. Coincidentally, just after Mobileye released the video, Jidu's intelligent driving technology solution was exposed for the first time, integrating pure vision and dual laser radar solutions into the car, and it is expected to achieve "true redundancy" after mass production. Previously, Xinna Sensing launched a new type of triple redundant sensor to escort autonomous driving safety with precise perception. At a time when Tesla's "pure vision" autonomous driving technology route has been criticized, will redundant sensing be the future of autonomous driving?

The “eyes” of autonomous vehicles

Safely and successfully bypassing pedestrians crossing the road and completing a U-turn at an intersection with multiple traffic lights... In the unedited video released by Mobileye, despite poor road lighting at night and complex road sign design, Mobileye's self-driving car still successfully completed complex driving operations.

The system that gives Mobileye's self-driving cars "perception superpowers" is a truly redundant sensing system that Mobileye says consists of two independent subsystems, one of which uses cameras and the other a combination of lidar and radar.

The vehicle is equipped with a normal working sensor and a backup sensor system. The redundant sensor system mentioned by Mobileye is actually like people's two eyes. In an interview with a reporter from China Electronics News, Gu Rongxiang, general manager of Xi'an Zhongxing Measurement and Control Co., Ltd., made such a vivid analogy: people's left eye and right eye can be assumed to be two image sensors. When one eye is blocked, the other eye can still distinguish objects normally, without affecting normal life.

Dong Bo, chief algorithm engineer at ACEINNA Sensing, took the triple redundant IMU (inertial measurement unit) module as an example to further explain to the reporter of China Electronics News how redundant sensing systems play a role in the field of autonomous driving. He said that traditionally, an IMU module contains a three-axis accelerometer and a three-axis gyroscope. When any axis of the accelerometer or any axis of the gyroscope fails, the entire IMU module will fail. ACEINNA Sensing's triple redundant IMU module integrates three three-axis accelerometers and three three-axis gyroscopes. The entire IMU module will fail only when more than one three-axis accelerometer or three-axis gyroscope fails.

In order to significantly improve the safety of high-level autonomous driving, redundant sensor systems are essential. Xu Tianyu, senior analyst at CCID Consulting's Internet of Things Industry Research Center, told China Electronics News that in the field of environmental perception for autonomous driving, ultrasonic radar, millimeter-wave radar, camera, and lidar will form at least dual 360-degree perception capabilities, that is, at least two sets of sensor systems are guaranteed, each of which can achieve 360-degree detection around the vehicle without blind spots. When one system fails, the other set of equipment can automatically switch and immediately take effect to ensure driving safety.

In general, based on safety considerations, the video sensors and millimeter-wave radar sensors of autonomous vehicles must be designed with redundancy, and multiple sensor solutions must be considered. Gu Rongxiang explained that this is because the road surface and driving environment are complex and changeable, and the driving state of the car itself is prone to sudden changes, so the front end of the autonomous driving system needs to "keep an eye on all directions and listen to all directions."

The "promotion test" that cannot be avoided

Currently, Mobileye's self-driving taxis equipped with "truly redundant sensing systems" are planned to be put into small-scale commercial use later this year; Jidu Auto Robotics is constantly moving forward on the road to high-level autonomous driving by adding redundant dual systems and dual lidars; previously, ACEINNA Sensing launched a new type of triple redundant sensor to safeguard the safety of autonomous driving.

Redundant sensor systems are being installed one after another, but before installing redundant sensors in self-driving vehicles, it is not easy to help the redundant sensors "find a place" and "position" them. Gu Rongxiang said that when installing redundant video sensors, the first thing to do is to comprehensively consider multiple factors such as direction, clarity, night driving, wind, rain, and sandstorms, and then determine the installation location of the sensor; when installing redundant millimeter-wave radar sensors, multiple factors such as direction, radar perception distance, and resolution must be considered.

The successful installation of redundant sensor systems on vehicles is not the "end" of the autonomous driving business. To achieve fully autonomous driving, redundant sensor systems must also go to the market and achieve large-scale promotion and application. However, it should be noted that due to the complexity of technology and high costs, redundant sensor systems are facing an inevitable "promotion test".

The first challenge is technology. Xu Tianyu said that in terms of sensor miniaturization, only when the sensor size is reduced can the installation of redundant sensor systems be guaranteed. For sensor miniaturization, MEMS technology can be used to integrate microsensors, microactuators, micromechanical structures, micropower sources, microenergy, interfaces, etc. on a single chip, greatly reducing the size of the sensor.

In terms of redundant system switching, it is necessary to consider how to judge faults, how to achieve switching without human intervention, uninterrupted switching between primary and standby systems, and no user perception. Xu Tianyu said that this requires the use of edge computing and edge controllers to achieve this.

其次是来自成本的考验。激光雷达每台近1000美元价格，考虑冗余设计后车辆成本将显著增加。对此，徐田雨表示，如果出于降低车辆成本考虑，也可以采用半导体制造方法(MEMS等)，批量制造以降低单个传感器成本。

In Gu Rongxiang's view, sensors are related to driving safety, so factors such as accuracy must be taken into account; sensors must have a very high safety level, and they need to comprehensively and fully consider external environmental factors, road environmental factors and the vehicle's own conditions, and accurately perceive in complex and changing internal and external environments. This is undoubtedly a huge technical challenge for sensors. In addition to improving performance parameters such as accuracy and intelligence, the reliability and environmental adaptability of the sensor system are also very important.

Sensor systems are developing towards vehicle-road collaboration

Looking ahead, is redundant sensing the only way to achieve autonomous driving? Dong Bo told the reporter of China Electronics News that multi-sensor fusion is a necessary condition, but whether redundant sensing is necessary depends on whether a single (non-redundant) IMU can meet the reliability requirements and the design of the entire autonomous driving system.

There is a difference between multi-sensor fusion and redundant sensing. Dong Bo explained to reporters that multi-sensor fusion is usually used to achieve functions that a single sensor cannot achieve, or to overcome the inherent deficiencies of a single sensor; redundant sensing means that when one or more sensors fail, the entire system can still work normally and the performance indicators are not affected.

In any case, to achieve true autonomous driving, "smart" cars need to develop in coordination with "intelligent" roads. At present, the increasing safety requirements for autonomous driving have prompted the sensor system to continue to develop in the direction of vehicle-road collaboration.

Why will vehicle-road collaboration become an inevitable trend in the development of sensor systems? In fact, this is because on-board sensors cannot fully and accurately perceive road conditions. Installing perception sensors on both sides of the road can provide a "God's perspective" for autonomous vehicles, complementing the information from on-board sensors.

Xu Tianyu further explained to the reporter of China Electronics News that the sensing range of vehicle-mounted sensors is actually quite limited. Taking millimeter-wave radar and laser radar as examples, the maximum detection range of the former is 200 meters, and the maximum detection range of the latter is 150 meters. However, when multiple roadbed sensors are connected to the Internet, their detection range can theoretically reach infinity, which can effectively expand the sensing range of autonomous vehicles.

In addition, vehicle-mounted sensors can only detect the environment but not themselves, while the application of roadbed sensors can effectively supplement the information from vehicle-mounted sensors.

The organic coordination of environmental sensors, road sensors and vehicle sensors can provide more reliable connections and accurate judgments for autonomous vehicles. Gu Rongxiang told reporters that installing corresponding sensors on both sides of the road can better solve some problems of autonomous driving perception, but this is only one aspect of solving the coordination between the vehicle and the environment, and cannot completely solve the problems of autonomous driving safety and reliability.