Nullmax Introduces New Vision Technologies to Boost Self-Driving Performance, Delivering Exceptional Mobility Experience

Nullmax, a pioneering leader in autonomous driving technology, is introducing a suite of new vision technologies for perception to enhance the autonomous driving capabilities. These technologies include dynamic and static scene understanding, such as 3D lane detection, 3D object detection, and real-time local map , aimed at delivering exceptional autonomous driving experience.

Perception is the upstream section and foundation of autonomous driving, with vision perception playing a crucial role. The quality of vision perception directly impacts the capabilities of autonomous driving systems, especially for enriching functionalities and achieving peak performance.

Globally, there is significant demand for mass-producible vision perception in self-driving, and few companies are capable of providing it. Nullmax stands out as one of these companies that offers mass-production-level vision perception capabilities and whole self-driving solutions across any computing platform. Based on the deep understanding and extensive experience in autonomous driving, Nullmax has developed a series of cutting-edge vision perception technologies.

In its quest to enhance the intelligent driving experience and achieve highly intelligent and autonomous Ultimate Self-Driving (USD) function, Nullmax has recently rolled out perception algorithms and methods: QAF2D, CurveFormer, and BEVSegFormer. These researches address the technical challenges of autonomous driving in complex urban environments and have been accepted at top academic conferences such as CVPR 2024, ICRA 2023, and WACV 2023.

Specifically, QAF2D introduces a novel multi-camera 3D object detection method that utilizes 2D detection to guide 3D detection, thereby improving obstacle detection performance in terms of stability and accuracy. CurveFormer directly calculates 3D lane parameters in Bird's-Eye-View (BEV) space in an efficient way, and BEVSegFormer allows vehicle to generate high-precision local map in real-time, enabling Navigate on Pilot feature on any road without the need for high-definition maps.

These advanced technologies not only demonstrate outstanding performance but also are ideally suited for deployment on current vehicle computers. They can be readily implemented in upcoming new vehicle models without the need for extravagant hardware setups. Furthermore, Nullmax continues to enhance these technologies, with upgraded versions progressively being completed.

Nullmax has incorporated these technologies into its MaxDrive solution, which provides customized applications for vehicles with varying configurations. The high configuration offers top-tier self-driving capabilities including USD function. This solution enables L3 and L4 levels of autonomous driving, featuring an end-to-end AI neural network capable of handling complex and unfamiliar scenarios, ensuring a remarkable user experience.

Currently, Nullmax is conducting road tests of these technologies and solutions and collaborating with international automotive giant to advance their commercial deployment. Preliminary achievements are expected to be seen in 2024.

Nullmax is dedicated to developing vision-centric autonomous driving solutions that are not only highly capable in complex road conditions but also cost-effective and deployable. These advanced vision technologies ensure that Nullmax's solutions remains a leader in safety, comfort, and affordability.

Nullmax believes that superior vision perception technologies can create the best autonomous driving systems, thus delivering consumers a novel technological experience with global partners.