ADI builds a high-performance ToF development platform from software and hardware to accelerate application implementation scenarios

Game players perform various actions in the air with their bare hands in front of the gaming devices, easily taking on various game characters in the virtual world; robots of all colors navigate autonomously in various scenarios, easily avoid obstacles and move quickly to achieve a variety of specific target tasks such as autonomous cleaning, express delivery or patrol; when truck drivers are drowsy in the afternoon, they stop to rest for safety when they hear constant warning sounds from the cockpit... These are all real-life technology application scenarios that are increasingly appearing in our lives, and ToF (time of flight) is becoming one of the key enabling technologies for these many innovative applications.

ToF has quickly entered the public eye from its initial trendy application in mobile phone cameras, and has begun to show its prowess in various application fields. According to the IHS Markit report, based on the many advantages of the ToF solution, the ToF market size is expected to reach US$1.5 billion in 2022, accounting for about 50% of the 3D sensing market. In a recent media interview, ADI Industrial Market Manager Li Jia also believed that as one of the three major solutions in the field of 3D deep vision, ToF technology is not only used in mobile phones, but also in VR/AR gesture interaction, automotive electronics ADAS, security monitoring, and new retail. It has begun to show its prowess in many fields, and is expected to promote a new wave of 3D intelligent sensing application innovation.

ADI Industrial Market Manager Li Jia

The ecosystem is booming, driving a new wave of 3D smart sensing applications for ToF technology

As we all know, the biggest advantage of ToF compared with binocular and structured light is distance measurement. Its depth calculation accuracy often does not change with distance, and can basically be maintained at the centimeter level. Therefore, in scenarios involving large-scale movements, ToF is very applicable. Before 2010, ToF technology was mostly used in scientific research fields such as cosmic measurement and high-precision microscopy. With the substantial improvement in the performance of light-emitting components, industries such as autonomous driving, VR/AR, industrial robots, and the Internet of Things have gradually emerged, further driving the development of the ToF application market. "For now, the application scenarios with the highest matching degree of ToF include ranging and perception of the driving environment in autonomous driving, monitoring of safe distances for human-machine collaboration in the industrial field, machine vision, volume and calculation in the logistics industry, robot navigation, etc., all of which fully utilize the advantages of ToF sensors in precise positioning and distance measurement." Li Jia pointed out.

Pulse ToF is one of the depth measurement technologies. In order to accurately synchronize light pulses, these calculations need to be performed on millions of pixels per second, and they must be adjusted according to working conditions. This is extremely challenging for mixed-signal circuit design and application. ADI is one of the few companies with technical expertise that can provide ToF solutions that are both high-performance and cost-effective. In 2014, ADI began to customize ToF technology for a well-known foreign AR glasses. The product was successfully launched and realized human-computer interaction and 3D reconstruction functions; in 2016-2017, ADI ToF technology was used for in-car gesture recognition; in 2018, ADI ToF technology began to cooperate with a domestic brand of mobile phones and was successfully launched in batches... "So far, ADI's ToF technology and products not only cover industries, consumers, and automobiles, but the boundary expansion of applications is still continuing. In the future, larger space, higher-resolution depth data combined with powerful classification algorithms and AI will unlock more new scenarios." Li Jia said.

ADI ToF Solution Application Development Timeline

In the past period of time, ToF applications have gradually shown an accelerated landing trend. For example, SLAM (Simultaneous Localization and Mapping) has begun to play an active role in the autonomous positioning and navigation of robots, effectively promoting the realization of human-computer interaction; the ToF module developed by ADI combines image sensors and VGA ToF sensor modules with built-in image processor solutions to provide a wider range of collision detection and prevention for applications such as car reversing systems, door opening protection systems, parking assistance systems and blind spot detection. "To further enhance the large-scale application of ToF technology in the industry, it is always necessary to cooperate with upstream, midstream and downstream industry chain partners, including optics, electricity, chip level, circuit level and module factories, and even some applications require professional teams to develop and adapt the entire solution." Li Jia pointed out. It is reported that ADI has launched a 3D time-of-flight development kit with Arrow, and has cooperated with several software solution companies such as Yuanjing Intelligence to launch a complete solution that meets a variety of application scenarios.

ADI ToF Industry Chain

Analysis of ADI ToF analog front-end core chip to build a precise measurement hardware platform

Generally speaking, ToF technology can be divided into two types according to the modulation method: pulse modulation and continuous wave modulation. Both ToF modulation methods have their own advantages and disadvantages. Points that need to be comprehensively considered according to the actual application use case include measurement distance, environment in which the system is used, accuracy requirements, thermal/power consumption limitations, size, and power supply issues. According to Li Jia, ADI's ToF solution currently mainly adopts the pulse principle of CCD sensors. As a global exposure device, CCD has very good performance in outdoor long-distance scenes. At the same time, CCD is a global exposure sensor and has narrow shutter exposure, which has very strong anti-interference performance against external interference.

Comparison of the performance of a CMOS ToF system (850 nm light source) and ADI's CCD ToF system in identifying people and tripods in an outdoor environment.

Specifically, ADI's ToF solution uses a high-performance ToF CCD and a ToF analog processing front end that integrates a 12-bit ADC, a depth processor (processes the raw analog image signal from the CCD into depth/pixel data), and a high-precision clock generator (generates drive timing for the CCD and laser). The ToF CCD main chip ADDI903x series has excellent performance. The sensor converts the collected light signal into an electrical signal, uses depth calculation to send the final distance information to the SOC for processing, and then provides the data to the algorithm adaptation. The entire process involves sensors, lenses, transmitting devices, platforms, applications, etc. For hardware manufacturers, they need to cooperate with module manufacturers to create small and highly integrated products to adapt to application manufacturers, and then adapt to different application scenarios.

Functions provided by ADI CCD ToF front-end chip ADDI903x

Since the ADI ToF system-level solution uses a ToF image sensor with a resolution of 640×480, its resolution is 4 times higher than that of most other ToF solutions on the market. At the same time, compared with CMOS solutions, this system-level ToF solution can provide higher system performance at the same size or cost. For example, high resolution can better distinguish the subject from the background in a complex lighting environment. Thanks to the CCD architecture designed for the 940nm light band, the ADI ToF solution can also capture images in a moving environment more accurately.

In addition to front-end chips and sensors, more optoelectronic devices are also required according to the needs of CCD, and the peripheral devices of ADI's ToF solution are all products of ADI. "ADI ToF system-level solution does not recommend customers to make too many modifications when using it, because each module in the system is the best overall performance state achieved by ADI and its partners after long-term running-in." Li Jia emphasized.

ADI ToF solution complete system composition

Find and solve potential technical pain points and explore broader application paths for ToF

More and more application scenarios have opened up new opportunities for ToF, but related challenges may also arise, such as how to avoid interference between multiple ToF application terminals in the same scenario. Especially for professional-level scenarios such as automobiles and industries, the design of ToF sensing systems not only needs to strike a balance between accuracy, range, response time, resolution, cost, power consumption and available packaging requirements, but also needs to target various uncontrollable factors that arise in different actual situations. Customized designs are required for the flexibility and anti-interference of the sensing system, such as adding some high-reliability filtering and anti-interference devices and modules, and loading related software algorithms to ensure that the system has sufficient capabilities to cope with different types of emergencies.

Imagine if multiple autonomous robots are sorting goods in the same large warehouse, or two self-driving cars are approaching an intersection at the same time, and the ToF camera cannot eliminate the interference of each other's light sources, then the application scope of using ToF technology for accurate depth measurement will be severely limited (ADI's development roadmap is expected to achieve a scenario where 64 cameras are simultaneously detecting at close range). Even on the consumer side, as various applications increase, interference between devices will be a real practical problem. Therefore, ADI believes that the ability to prevent or eliminate interference in ToF systems will become increasingly important. It is revealed that ADI currently uses a patent-pending algorithm that can avoid or eliminate all irrelevant light information and only use the light information of its own laser source, so it can give correct depth information. The depth processor uses a pseudo-randomization algorithm and special image processing functions to eliminate multi-machine interference. Therefore, multiple ADI ToF systems can be used in the same environment.