Autonomous Driving Systems

3 Questions for… Dr. Liming Chen, Horizon Robotics

“The industry is trending toward deeper hardware and software integration“

1 min
Dr. Liming Chen will be one of the keynote speakers at 2025 Automobil-Elektronik Kongress.
Dr. Liming Chen will be one of the keynote speakers at 2025 Automobil-Elektronik Kongress.

Ahead of the Automobil-Elektronik Kongress 25, we talked to Dr. Liming Chen, President at Horizon Robotics, about the latest trends in the field of advanced driver-assistance systems.

Dr. Chen, what new paradigms are emerging in the ADAS automotive industry?

China has become a global center for consumer-driven innovation in smart mobility. OEMs and technology companies in China are moving fast and boldly, typically releasing new vehicle models every 1.3 years (according to a report by AlixPartners in July 2023), making the region a proving ground for new ideas. Meanwhile, software is evolving rapidly. End-to-End Transformers are being deployed, and LLMs/VLMs are now being integrated into the driving stack, yielding safer and more human-like driving experiences. The industry is also trending toward deeper hardware and software integration to streamline the complexity of current systems. We anticipate the emergence of a unified AI agent running on an integrated processor within the next 3 to 5 years.

Automobil-Elektronik Kongress 25

  • The summit of decision-makers in automotive electronics
  • On 24th and 25th June 2025 at the Forum Ludwigsburg
  • VIP event by personal invitation on 23rd June 2025
  • At the evening industry gathering on 24th June, there are excellent networking opportunities.
  • Info and registration: www.automobil-elektronik-kongress.de

How is Horizon Robotics leveraging advanced algorithms to fuel progress in ADAS technologies? 

At Horizon Robotics, our deep-rooted expertise in software is central to our mission to make human life safer and better through smart-driving vehicles. We are recognized as an industry leader with acclaimed algorithms such as UniAD, Sparse4D, Vision Mamba and Senna. Our team is pioneering the development of End-to-End models and Vision-Language Models. Our approach integrates software innovation with purpose-built hardware through a tight co-design. This includes our proprietary ML accelerator, the BPU, engineered for high efficiency and tailored for smart mobility. The result is a new generation of ADAS and autonomous driving solutions that are not only safer and more responsive, but also intuitive and consumer-centric.

What are the key challenges in deploying ADAS solutions at scale? 

To advance ADAS and autonomous driving technologies, solutions must scale in two dimensions: performance (scale-up) and operational design domain (scale-out). Success depends on being user-centric, accelerating deployment, and achieving scale across all vehicle price ranges. Meeting these demands requires agility in adapting to evolving consumer expectations, managing compute cost and energy efficiency, and mastering the complexity of modern software stacks—all while upholding the highest standards of safety and security. The road ahead presents significant technical and operational challenges. Sustained investment in both hardware and software is essential, as is deep collaboration with partners across industries and regions. Only through this integrated approach can we unlock the full potential of smart mobility!