CICV 2023 | "Smart Vehicle Human-Computer Interaction Technology" seminar was successfully held

In recent years, the continuous innovation of human-computer interaction and artificial intelligence technology has provided richer application scenarios for intelligent connected cars and provided drivers and passengers with a safer and more pleasant travel experience, which will accelerate the development of intelligent connected cars. Industrialization.

With the 2023 China (Yizhuang) Intelligent Connected Vehicle Technology Week and the 10th International Intelligent Connected Vehicle Technology Annual Conference being held in Beijing, the "Smart Vehicle Human-Computer Interaction Technology" seminar on May 17 was also held by This meeting was successfully held. The meeting focused on sharing and communicating on new technologies such as user experience, vehicle-machine interconnection, human-machine hybrid intelligence, and new AI-driven interaction platforms.

This meeting was chaired by Professor Shen Chuanliang of Jilin University, Professor Song Xiaolin of Hunan University, Assistant Professor Lu Chen of Nanyang Technological University, Deputy General Manager of iFlytek Zhang Wei, OPPO Strategic Planning Manager Guo Yi, Shen Chuanliang, a professor at Jilin University, and Hu Chuan, an associate professor at Shanghai Jiao Tong University, respectively delivered keynote speeches for this conference. Many audiences from universities, enterprises and industry research institutions across the country attended this conference.

Song Xiaolin, a professor at the School of Mechanical and Vehicle Engineering of Hunan University, published a keynote report on "Human-machine shared control and multi-dimensional evaluation of autonomous driving systems considering interactive behaviors", sharing the sharing of driver interactive behaviors Control strategy design and human-machine control authority allocation strategy design based on driving safety field and game theory. He said: Judging from the research results, the degree of intervention and driving risk are lower than if not considered, which shows that the control is effective and the collaborative mechanism is effective. At the same time, the "User Adaptability Assessment Model for Autonomous Driving System Roles" was introduced from three aspects: description of takeover criticality, changes in trust, and construction of driver trust model in human-machine collaboration. He said that as the number of takeovers increases, the degree of risk will increase after the system is proficient. Will decrease, the longer the takeover time is, the trust level will also decrease.

Nanyang Assistant Professor Lu Chen of the School of Mechanical and Aerospace Engineering at Nanyang Technological University in Singapore delivered a video speech on "Human-machine hybrid intelligence for future transportation", focusing on human-machine collaborative systems and human-like autonomous systems. Driving aspects were introduced, and the human-machine collaborative two-way enhanced control algorithm architecture and shared control algorithm were shared. Under the framework of the human-machine system, multiple supporting sub-modules were developed at the bottom, using machine vision technology to build depth The model monitors the human driver's activities in the cockpit. Regarding the realization of human-like machine intelligence, this paper introduces how to use reinforcement learning-like methods to reconstruct and model human driving behavior. He said: He hopes that the results of future interactive predictions can be improved more accurately to further help downstream decision-making and control modules, making decision-making road trajectory planning and control safer, more efficient, and more robust.

Zhang Wei, deputy general manager of the intelligent interactive product line of iFlytek's Intelligent Automotive Division, delivered a speech on "New Generation Interactive Cloud Platform for Automotive Scenarios" and introduced how to define a new generation of interactive cloud platform for automobiles. He said: The first is full-link low-code and zero-code definition of brand specifications, which can realize low-code development; the second is the integration of deep understanding capabilities of large models and the continuous purification and upgrade of multi-modal interactions, so that skills and line networking can Achieve seamless penetration; the third is openness, which can realize unified management of multiple types of excellent AI capabilities and new ecology on the platform side, and aggregate scheduling; the fourth is high efficiency, flexible definition and organization of full-link dialogue interactions, and car cloud Collaboration helps product differentiation and innovation. Based on the above four points, iFlytek introduced the product "Car Interaction Cloud Platform CIVI", a one-stop intelligent assistant product development solution that integrates the Spark cognitive large model and artificial intelligence development technology to realize the needs of in-vehicle intelligent assistant products. Capability aggregation, differentiated interaction, sustainable evolution and iteration and other service functions

OPPO Senior Technology Strategic Planning Manager Guo Yi delivered a speech on "The Future of Integration of Mobile Phones and Cars", detailing the current status of the interconnection of mobile phones and cars, the possible integration of mobile phones and cars, and the future of mobile phones and cars. future ecology. He said: Mobile phones are the extension of people, and cars are the extension of space. They are dependent on and complementary to each other. Car and mobile phone users, usage scenarios and usage durations are highly overlapping. The degree of fragmentation of the software and hardware of the car itself determines that the current ecological construction will still take some time. In the future, mobile phones and cars will be more integrated. In the car interface, you can see that the mobile phone will supplement the ecology of the car. The second user will use the services of the mobile phone and the car at the same time. Through the integration solution, the ecological shortcomings of the mobile phone and the car will be filled.

Professor Shen Chuanliang of the School of Automotive Engineering of Jilin University delivered a speech on "Research on Tactile Feedback Technology in Intelligent Cockpits", covering the modalities of human-vehicle interactive information, the mechanism of tactile perception, and the mechanism of tactile feedback. , the mechanism of tactile interaction form, autonomous driving tactile feedback, tactile feedback research formation, shared the current status of tactile feedback technology in detail, and also put forward the following prospects for future technology development, vigorously developing active tactile feedback systems; realized through acoustic wave technology Non-contact tactile interaction; develop three-dimensional haptics on smart surfaces to reduce visual occupation and improve safety.

Hu Chuan, associate professor at the School of Mechanical and Power Engineering of Shanghai Jiao Tong University, published "Research on the Quantitative Model of Trust in Human-Machine Public Driving", which was based on the following trust degree modeling, objective trust degree modeling and Several aspects of trust control were shared in detail. For example, to evaluate trust, the driver's physiological indicators such as skin electrode, ECG, heart rate, eye movements, etc. are evaluated, and the driver is asked to score questionnaires. The physiological indicators and the driver's perceived risk score are combined to extract feature definitions and use deep learning. The method models perceived risk and estimates trust in real time. In addition, the external environment will also affect the degree of trust, so the real-time estimation of trust is transmitted to the trust bias calculation framework.

The above is the detailed content of CICV 2023 | "Smart Vehicle Human-Computer Interaction Technology" seminar was successfully held. For more information, please follow other related articles on the PHP Chinese website!