TBC
China’s digital development is driving AI-enabled automotive experiences (multimodal interaction). From focusing on technical functionality to satisfying user emotional value, technology fosters security and trust, and new technological architectures enable the exploration of new business models. Also, the differences between Chinese and international users will be briefly analyzed.
Regulation changes in China set ADAS Industry under big pressure – what does it mean?
This sharing focuses on the IAV & cogniBIT advanced traffic flow simulation solution, with the core goal of verifying the cognitive AI driver model through a series of steps. Firstly, it processes the real driving data, converts it into the specified format, and mines the model parameters. Then, it selects key scenarios to support simulation. Subsequently, it converts and analyzes the simulation data after integrating these AI drivers, compares and verifies the authenticity of drivers’ macro behaviors, and iteratively optimizes the parameters. Finally, it delivers reports showing the high simulation fidelity.
Elmos Semiconductor, a global leader with over 25 years of expertise in automotive ultrasonic sensing, continues to set industry benchmarks through innovation, performance, and market-driven design. With a strong focus on the Chinese automotive sector, Elmos offers tailored solutions that meet regional requirements while maintaining global excellence. Our proven 524.17 DSI3 product has long dominated the AK2 market in Asia. Now, its successor the 524.20 delivers greater flexibility, cost-efficiency, and soon, a raw data interface enabling advanced AI-based perception. Validated by leading OEMs and Tier-1 suppliers worldwide, this platform empowers next-generation ADAS and autonomous parking systems with unparalleled precision and reliability. The Elmos NextGen Ultrasonic Solution integrates advanced signal processing, high-speed digital communication, efficient power management, and robust cybersecurity into a single platform. Featuring built-in AI-powered Height Classification and ultra near-field measurement algorithms, it is purpose-designed for autonomous parking offering cost-optimized architecture, localization flexibility, and seamless integration for both premium and mass-market vehicles. Looking ahead, Elmos is preparing to introduce the DSI4 standard, redefining communication capabilities and setting a new benchmark in ultrasonic sensing for the autonomous era.
From ADAS to Smart Tier 0.5: How China’s Digital Ecosystems are Re-Shaping Intelligent Connectivity – China is leading the world in the rapid deployment of advanced driver assistance systems (ADAS) and intelligent connected vehicles. This transformation is powered not only by traditional suppliers but also by digital ecosystems that are increasingly shaping the mobility landscape. These ecosystems—rooted in data platforms, AI, and connectivity—are driving the emergence of a new “Smart Tier 0.5” supply chain, bridging the gap between automakers and end-users. In this presentation, Bill Russo will share insights and examples from the ADAS Guide, illustrating how China’s unique innovation environment accelerates the transition toward intelligent connectivity. He will highlight the competitive dynamics, implications for global suppliers, and the strategic role of cross-industry collaboration in defining the next era of in-cabin intelligence.
TBC
TBC
TBC
SDV and Autonomous Driving are the trend of automotive industry. They are facing many challenges, such as software complexity, security, mission critical and safety, software life cycle management, etc. This presentation will address these challenges, provide some solutions.
A Futuristic Vision on a RADAR System for Automated Driving – The RADAR sensors are today the most important complement to cameras for assisted and automated driving. Their active measurement principle allows direct 3D measurements with additional velocities. Combined with high robustness to weather influences makes them the key redundant modality each safety system is required to have. However classical RADAR sensors with their sparse point clouds are not as performant as LIDAR sensors. Imagine a network of cognitive digital broadband RADARs with hundreds of channels with large apertures spanning around the entire vehicle, cooperatively combining all channels of spectrum data at a central brain with modern AI methods – such a RADAR system would outperform a LIDAR system. Despite such an ultimate RADAR setup will take its time to get into the vehicles in all parts, it is showing the different technology paths industry is heading to. The talk will introduce the different technologies and show their benefits and the open challenges to be solved before realizing this ultimate RADAR vision.
China has become the best fitness centre for smart driving. We will talk about what’s the key factors to win in this market and how to accelerate the development and deployment in smart driving. The unique way of Horizon Robotics’ software and hardware co-optimization best practice shows the way from L2+ to L4/5. We will also show our current progress and products about smart driving and the key technology behind it.
TBC
TBC
High-resolution automotive radar with 6T8R or 8T8R is popular for OEMs in China due to its high performance and acceptable cost. Traditional way of building this kind of radar device is cascading two transceiver chips together and connecting them onto a strong radar MCU. Some new solution directly uses a newly designed 8T8R transceiver to replace a cascading method. We implement the system using a different way: directly cascade two radar SoC chips together. Besides cascading the LO signal, we connect two SoC chips by a SerDes interface. So, each SoC can access the other one using memory map method. All the resources of two chips can be utilized for processing. Some jobs are done by distributed processing, and the others are done by gathering required data in a single chip through SerDes transmission. In the talk, we will give the system architecture, an example of signal processing flow, the benefits of this kind of system. And we will give test results under different scenarios which are collected from some OEMs to show the performance of this system.
TBC
TBC
The automotive industry is undergoing a profound transformation, shifting from handcrafted vehicles to software-defined cars and now moving rapidly toward AI-defined mobility. While software-defined architectures (SDV) have enabled modularity, OTA updates, and continuous feature evolution, the transition has not been seamless. OEMs underestimated the complexity of scalable software integration, zonal E/E architectures, and the need for robust cloud-to-car ecosystems. Legacy development models, siloed supplier strategies, and slow organizational adaptation have often delayed innovation. Now, as AI-native vehicles emerge—with perception-driven ADAS, predictive maintenance, personalized in-cabin experiences, and self-learning digital twins—the gap between traditional OEMs and tech-driven players is widening. To remain competitive, the industry must embrace data-centric engineering, end-to-end software platforms, and AI-first development pipelines. This presentation explores the lessons missed, strategic shifts required, and how AI-defined architectures will fundamentally reshape vehicle design, business models, and the global automotive value chain.
TBC
InCabin monitoring systems require more and more functions every year. They become more complex and need to use more sophisticated and powerful sensors to comply with regulations and the beloved user ratings like Euro NCAP. How can you adapt to this ever-changing landscape making sure that the systems, not only work, but that they are robust enough to corner cases that yield false positives that reduce user acceptance? Learn how we can leverage the latest advancements in generative AI to complement physically based synthetic data to have full coverage of test cases with enough variability to make systems close to 100% reliable.
With the automotive industry’s rapid shift towards higher autonomy and enhanced safety, radar sensing remains crucial for reliable perception in all weather and lighting conditions. This collaborative presentation by GlobalFoundries and Arbe will delve into the synergy between semiconductor innovations and advanced radar system design. We will highlight how our joint efforts on the GlobalFoundries platform are pushing the boundaries of automotive radar technology. The presentation will cover semiconductor technology trends shaping the future of automotive sensing, the evolution of mainstream sensor technologies, and how these demands are driving innovation at the foundry level. Arbe will provide a deep technical dive into their latest automotive radar sensor, showcasing architectural choices, calibration and validation methodologies, and how the radar system addresses challenges such as mutual interference mitigation, ghost target suppression, and environmental robustness.
As Chinese sensor companies expand globally, success requires more than cost or specs—it demands trust, compliance, and strategic alignment. Drawing from his experience leading RoboSense’s LiDAR production program with Lucid Motors in the U.S., Peipei Zhao will explore four critical challenges Chinese firms face: system validation, regulatory hurdles, cultural differences, and go-to-market strategy. This session also offers guidance for global OEMs and Tier 1s: how to evaluate and collaborate with fast-scaling Chinese tech suppliers, and how to align processes across different engineering cultures. Attendees will gain a real-world framework for navigating cross-border collaboration in the next decade of global sensing innovation.
Details TBA
TBC
TBC
With the rapid advancement of automotive intelligence and connectivity, ultra-wideband (UWB) technology has evolved from positioning tools to core components that enable automotive intelligent sensing and secure interaction. Calterah’s first-generation UWB SoC series, Dubhe, has come into being. We have leveraged our over ten years of radar expertise to enhance the UWB radar capabilities of Dubhe, enabling it to be the World’s first 2T4R automotive UWB SoCs complaint with the latest standard IEEE 802.15.4ab. Combining outstanding RF performance, strong penetration capability and nanosecond-level pulse’s anti-interference performance, the Dubhe series of UWB SoCs will be a tech disruptor for in-cabin detection and thrive in current competitive and fast-moving automotive market. This keynote will elaborate on:
TBC
Leveraging the world’s data is now widely recognized as a cornerstone of AI advancement. Foundation models such as LLMs have been trained on massive datasets—often encompassing the entire Internet—to achieve groundbreaking results. In the domain of autonomous driving, teams face a similar demand for vast and representative datasets, capturing the world through each vehicle’s unique sensing modalities.
First-generation annotation pipelines raced to construct large-scale manual labeling workflows. Second-generation pipelines introduced offline AI-based pre-labeling to boost efficiency—but often at the cost of human adaptability and oversight.
In this talk, I will present a vision for third-generation pipelines: a human-AI fusion approach that combines the efficiency of automation with the flexibility and insight of scalable human-in-the-loop intelligence. The resulting pipeline is designed to be more efficient, adaptable, and robust—powering the next wave of progress in AD/ADAS data development.
TBC
TBC
