High-quality sensors for autonomous vehicles

Source: ASM Pacific Technology Ltd.

Compared with traditional cars, self-driving cars have obvious advantages

According to the Association for Safe International Road Travel ( ASIRT ), more than 1 million people die in road traffic accidents each year, most of which are caused by drivers. Self-driving cars are based on machine learning and computer technology, do not require human operation, and eliminate human errors such as drunk driving, speeding and fatigue driving, improve the safety of passengers and other road users, and reduce economic or casualty losses caused by traffic accidents.

此外，自动驾驶还可以增强驾驶体验，细致的速度调整使行车时更平稳，路线优化提供更省时路线等等，让乘客可以利用旅程时间在安全舒适的环境中，作其他活动，如阅读、工作、看电影或甚至休憩睡觉，进一步提高乘客的生活质素。

On-board sensors are the "five senses" of autonomous driving

There are two main core elements in the autonomous driving system:

1) Artificial Intelligence ( AI ) is the “brain” responsible for receiving, processing data and issuing appropriate instructions;

2) Advanced Driver Assistance System ( ADAS ) is the "five senses" of the car, responsible for inspecting the surrounding objects and environment, and transmitting information to the "brain" system, so that the latter can issue corresponding action instructions to the car's "limbs" to make the car perform corresponding actions.

ADAS is composed of a complete set of sensors, including short-range and long-range radars, LiDAR and precision cameras. Each sensor has a different perception method, receiving information outside the car in real time, providing the car with a complete environmental view. After the information is analyzed by the AI ​​brain, ADAS can identify different key categories, including pedestrians, obstacles, traffic signs, distance and speed of other cars, etc.

After comprehensive analysis of the data, the car is quickly "commanded" to respond appropriately, such as accelerating, decelerating, braking, steering, etc. As for the "command", it is a specific software program that gives specific instructions based on traffic regulations and operation modes, so that the vehicle can respond appropriately.

Zero-defect sensor production is key to enabling self-driving cars

According to a report by Yole Développement (see Figure 4), the compound annual growth rate of sensors will be 21% from 2020 to 2025, and the number of on-board cameras will increase from two to more than ten in thirteen years (see Figure 5), which shows that the autonomous driving car market is growing rapidly.

CamSpector - 3D Automated Optical Inspection for Smart Factory 4.0

oEasier and faster – the system can quickly identify elements such as “ dots ” and “ lines ” without the need for operator programming

o False positive rate can be reduced to less than 0.1 % and false negative rate can be maintained at 0% – the high-definition optical components included with CamSpector greatly enhance the system’s inspection performance

Other key features:

oEquipped with a dust filter, it can effectively remove 99.99% of pollutants as small as 0.1µm in diameter , keeping the dust inside the machine below level 100

o Fixed optical lens design and mobile workbench with magnetic rollers further eliminate the chance of re-contamination after inspection

o Its patented foreign body removal module uses more than five times more sticky gel sticks than other brands, helping to reduce consumables costs

Figure 6: ASMPT Automated Optical Inspection (AOI) System - CamSpector