World's first! RoboSense (02498) full-stack self-developed digital LiDAR chip passes AEC-Q certification, leading the technology in the industry.

RoboSense ROBOSENSE (02498) announced that both of its core chips for digital LiDAR have passed the AEC-Q102 automotive grade reliability certification, making it the only technology company globally to achieve automotive-grade standards for the entire chain of self-developed chips for digital LiDAR emission, reception, and processing. The self-developed SPAD-SoC chip and 2D addressable 2D VCSEL chip by ROBOSENSE have passed automotive grade certification, not only indicating the maturity of the new generation LiDAR chip technology developed by ROBOSENSE, but also marking the first time that core chips for digital LiDAR have completed automotive grade certification globally. This time, ROBOSENSE's self-developed SPAD-SoC receiving and processing chip (left) and 2D addressable 2D-VCSEL chip (right) have both passed AEC-Q series automotive grade certification. The AEC-Q certification, as a globally authoritative criterion in the automotive electronics field, was established by the Automotive Electronics Council (AEC) with Chrysler, Ford, and General Motors to create a universal set of component qualification and quality system standards. It verifies product quality and reliability through over a hundred tests including extreme high and low temperatures, temperature cycling, accelerated life, mechanical shock, EMC, etc. The testing standards are stringent, such as requiring extreme temperature tests at -40C and 125C to ensure the long-term stable operation of the chips under various harsh conditions. Traditional analog LiDAR requires seven key devices to complete emission, reception, and processing, while ROBOSENSE's innovative digital architecture only requires two core chips to achieve the same functionality. Under the digital architecture, the performance of LiDAR in terms of lines can exceed that of analog LiDAR by more than 10 times. ROBOSENSE's SPAD-SoC is the world's first chip to integrate SPAD receiving array and data processing SoC, meeting AEC-Q certification and realizing mass production application. In terms of performance, this chip is equipped with a super WAVE Engine, featuring a quad-core APU and dual-core MPU, capable of increasing the effective target recognition rate by 99% and supporting a high sampling rate of 440 billion times per second to ensure more complete signal acquisition and prevent missed detections. It also has a built-in 384-Core signal processing core, which efficiently processes signals in parallel, while maintaining a 0.25cm echo output resolution, three times higher than the industry standard. In terms of signal processing, it has a wide dynamic response range, can effectively detect low-reflective objects, and can withstand rain, fog, dust, and other environmental pollution, enabling precise differentiation and detection of objects in complex scenes and overcoming the limitations of short-range blind spots. In terms of packaging technology, ROBOSENSE is the first globally to integrate SPAD and SoC on a single chip through 3D stacking integration and using copper-copper bonding technology, providing each SPAD with independent processing circuitry, ensuring lossless and efficient signal transmission, and reducing the volume of LiDAR systems by over 50%. ROBOSENSE's 2D VCSEL chip is the world's only mass-produced 2D addressable VCSEL emission chip. This chip uses 2D addressable array scanning technology to successfully suppress high reflection expansion and solve the world-class industry problem of single-photon crosstalk. Moreover, the 2D VCSEL chip supports flexible matrix partition scanning, further boosting laser emission efficiency with the highest energy utilization rate of up to 70%, while simultaneously reducing power consumption and being more friendly to heat dissipation. In terms of integration technology, the ROBOSENSE 2D-VCSEL chip is highly integrated, with a dual-core SiP co-packaged design that combines the VCSEL laser array with the driver unit, significantly reducing the PCB footprint by 70% and greatly reducing the parasitic parameters of the system. The chip can be directly interconnected on the substrate, facilitating the performance improvement of the emission circuit, stable laser pulse width control within 2 nanoseconds compared to the previous 5 nanoseconds, and a substantial increase in ranging accuracy. Behind the leading performance advantages of the self-developed chips are the support of ROBOSENSE's three original patented technologies: Anti-crosstalk encoding technology, Crosstalk Immunity Encoding (CIE), encodes every transmission signal independently, overcoming environmental interference such as rain, fog, and intense light, resulting in a 95% increase in sunlight noise resistance and a 90% increase in crosstalk interference resistance. Picosecond precision jitter technology, PicoSecond Timing Technology (PTT), which increases the deviation of each transmission signal from the nanosecond level to nearly a thousand times, achieving picosecond-level phase accuracy, increasing ranging accuracy by 15 times to 3cm, and simultaneously reducing power consumption. Adaptive fusion algorithm, Adaptive Fusion Algorithm (AFA), which adaptively extracts the echoes, expands the LiDAR measurement range, and selects the optimal part of the optimized signal, reducing the blind spot by 70 times. ROBOSENSE's original patented technologies and core chips collectively build the technological barriers for digital LiDAR. Compared to traditional analog LiDAR, ROBOSENSE digital LiDAR has advantages such as high resolution, strong anti-interference ability, and small size, leading the product performance. Currently, based on a full-stack chip architecture, ROBOSENSE's E platform and EM platform digital LiDAR matrix have fully entered the automotive grade production phase, covering a wide range of applications including Robotaxi, ADAS, and services for Siasun Robot & Automation, meeting the demands for technology upgrades. As a benchmark product in the LiDAR matrix, ROBOSENSE's high-performance LiDAR EM4 combined with E1 has become the preferred choice for L3 and L4 autonomous driving. To date, EM4 has won the main LiDAR spot for multiple leading Robotaxi companies in North America and domestically, while E1 has become the preferred solution for solid-state supplemental blind spot LiDAR in Robotaxi scale operations with eight major global companies having officially designated it. Furthermore, the "strongest combination" of EM4+E1 is becoming the preferred configuration for the next generation of Robotaxi. The 7th generation model of Pony.ai is equipped with 4 E1 LiDARs for large-scale road testing, and the jointly developed Robotaxi by DiDi and GAC Aion utilizes 6 E1 LiDARs to build a nearby perception protection network. The EM platform series products have accumulated 8 OEMs with placements on 45 models, among which the true 192-line EMX has secured sole supply placements for up to 32 models of a leading global new energy vehicle manufacturer. Currently, the Jek-9X and Zhi Ji LS6 vehicles equipped with the ROBOSENSE EM platform customized 520-line digital LiDAR have been launched and delivered to the market. From level 2 assisted driving to level 4 autonomous driving, ROBOSENSE will continue to lead in LiDAR chip technology innovation, providing high-quality, high-performance, and safe and reliable high-precision sensing for driving safety across all scenarios.