Supported by National Science and Technology Council (NSTC), Taiwan, the interdisciplinary research team, led by Associate Professor Yu-Chen Lin at Feng Chia University (FCU), has successfully developed the intelligent electronic suspen-sion system (IES) for improving the ride comfort and vehicle stability when driving on the pothole, manhole cover, puddle, expansion joint, and other road defects, by integrating different domain’s’ technologies, including artificial intelligence per-ception, embedded system, and interdisciplinary technologies. The developed IES can instantly identify the road defects in front of the vehicle, so that the suspen-sion system can be dynamically adjusted according to different road conditions, which not only effectively improves the ride comfort of the vehicle, but also in-creases the vehicle stability. The innovation system can not only promote the tech-nological upgrading of the domestic automotive electronics industry, but also en-ter the automotive electronics supply chain through localization.
Linking domestic industries to Translate a Real and High-Added Product: Intelligent Electronic Suspension System
Under Professor's Yu-Chen Lin leadership, the self-driving car team consists of six professors from Feng Chia University, National Taipei University of Technology, and Chung Yuan University as well as working with five automotive industries (Chimei Motor Electronics Co., Ltd., Trillion Green Energy Co., Ltd., Pin Hsiu Tech-nology Co., Ltd., Cybernet Systems Taiwan Co., Ltd., KAIFA Industry Co., Ltd.) to jointly develop a domestic intelligent electronic suspension system. In this re-search, the pavement defect detection system based on the deep learning tech-nology is proposed to sense the road surface conditions, including pothole, bump, manhole cover, expansion joint, puddle, water pothole, which effective recogni-tion distance is 20 meters for speeds below 100km/hr. Moreover, the perception system can identify the effective distance of 20 meters, and is suitable for vehicle speeds below 100km/hr. It is also implemented in the automotive embedded sys-tem, which the computing speed is greater than 20FPS (single core).
According the perception results, the deep reinforcement learning control strategy is then utilized to determine optimal control parameters for active sus-pension system in real time so as to ensure good ride quality; and further, it can reduce the risk of damage to your car, and increase vehicle stability when vehicle is going through a large puddle. Hence, automatic prediction and adjustment can be effectively implemented by pre-learning strategy. The components and controllers of IES are completely independently developed, which not only effectively en-hances the competitive advantage of Taiwan's automotive electronics industry, but also provides more outstanding riding comfort and vehicle safety of self-driving cars.
Critical Technologies Development for Opening Up a New Blue Ocean Market
The hype around self-driving cars has reached a fever pitch, with most indus-tries or researchers focused on the immediate innovation and novelty of driver-less. According to Gartner, Inc. , by 2023, worldwide net additions of vehicles equipped with hardware that could enable autonomous driving without human supervision will reach 745,000 units. American Planning Association (APA). In ad-dition, according to American Planning Association (APA), the transport policy dis-course posits Shared Autonomous Vehicles (SAVs) as a more sustainable solution for the implementation of road automation technology, especially for mass transit and freight markets. Therefore, in addition to vehicle safety, the ride comfort is an important issue of self-driving car. Nowadays, the development of autonomous car still focuses on safety and self-driving technology, without considering ride comfort. The developed IES not only has the advantages of accurate defect recog-nition, real-time suspension control, and price, etc., but also implements AI tech-nology into the automotive embedded system, and function verification. We be-lieve that the IES is expected to enhance the technology upgrade of domestic car electric suppliers, as well as provide new applications and innovative value for self-driving cars, opening up a new blue ocean market.
Road quality is a key factor in ride comfort, which affects vehicle wear and tear as well as safety in use. A new survey from American Automobile Association (AAA) pointed out that last year 1 in 10 drivers sustained vehicle damage signifi-cant enough to warrant a repair after hitting a pothole. With an average price tag of almost $600 per repair, damage caused by potholes cost drivers a staggering $26.5 billion in 2021 alone. The survey of Warranty Direct shown that the connec-tion between damage and potholed roads could be a key factor in up to a third of all mechanical failures in the UK, especially for suspension related problems ac-count for 42% of faults. In Taiwan, there are many types of underground pipelines, which often cause uneven road surface and even potholes or water accumulation due to repeated backfilling without stopping the digging of pipelines. This is not only an inconvenience for road users, but also for vehicles or self-driving cars, which can cause discomfort, damage to shock absorbers, or even a tire blowout and deviate from the lane.
It is noteworthy that the first dataset for Taiwan pavement defects (Taiwan Pavement Defect Image Dataset, TPDID) has been established, which the total number of TPDID samples is 89,723. Compared with the existing international road defect dataset, the advantages of the established TPDID dataset have more samples, multiple types of road defects, and conforming to the actual road condi-tions in Taiwan, such as the manhole cover on Taiwan's roads. The TPDID dataset can not only improve the accuracy of the developed AI-based road defects percep-tion system, but also effectively increase road repair efficiency by sending the road defect detection results to the roadway maintenance unit in real time.
Great Market Potential in Enhancing Safety and Comfort for Self-Driving Cars
Nowadays, most of the vehicle suspension systems are composed of tradi-tional rigid spring and fixed damper. Only a few high-priced top vehicles have an active suspension system composed of air springs or adjustable dampers; howev-er, it is still limited to manual adjustment of fixed segments (normal mode, com-fort mode, sport mode), and there is no self-adaptive adjustment according to dif-ferent road defects. By combining AI perception, decision and control technolo-gies, a novel intelligent electronic suspension system has successfully implement-ed, which not only promote ride comfort and vehicle stability when passing the road defects by adaptive adjustment of active suspension system in advance, and increase its durability. The relevant research results have not only been published in international journal papers [2-5], but also completed the integration of auto-motive embedded systems and installed on a commercial vehicle as well as func-tional testing. The price of the developed intelligent electronic suspension system is 1/3 of that of similar foreign products, which is very competitive in the automo-tive market. We also plan to cooperate with public transport operators, roadway maintenance units, smart manufacturers, and self-driving car operators to cus-tomize products according to different application requirements.
The lightweight artificial intelligence technology and embedded system have entered the automotive supply chain together with our partner, Chimei Mo-tor Electronics Co., Ltd., which has been actually installed in Nissan and other commercial vehicles, and the production value is expected to reach NTD 750 mil-lion. In addition, the developed AI perception and decision-making control tech-nologies have spawned a number of industry-university cooperation projects, such as Whetron Electronics Co. Ltd., TungThih Elecrtonic Co. Ltd, Foxtron Inc., Tong Ying Motor Co. Ltd., King Lung Auto Manufacturing Co. Ltd., etc., which the total amount of its derivative projects exceeds NTD 6.4 million.
Benefiting from the passing of ADAS regulations in the U.S. and EU in recent years, the developed intelligent electronic suspension system will create the more commercial opportunities for domestic vehicle suspension system manufacturers, automotive component suppliers, ADAS companies and even domestic chip de-signers etc. The developed intelligent electronic suspension system can be demon-strated to not only effectively improve ride comfort and robustness of autono-mous car, but also successfully upgrade their technological capabilities of the do-mestic automotive industry, as well as promote Taiwan's automotive electronics industry to enter the Tier 1 global supply chain, letting the world know Made in Taiwan.
. Meghan Rimol, “Gartner Forecasts More Than 740,000 Autonomous-Ready Vehicles to Be Added to Global Market in 2023,” Gartner, Inc., Nov. 14, 2019.
. Yu-Chen Lin, Ha Ly Thi Nguyen, Ji-Fan Yang, Hung-Jui Chiou “A Rein-forcement Learning Backstepping-Based Control Design for a Full Vehicle Active Macpherson Suspension System,” IET Control Theory & Applica-tions, vol.16, pp. 1417-1430, 2022.
. Yu-Chen Lin, Wen-Hui Chem, Cheng-Hsuan Kuo, “Implementation of Pavement Defect Detection System on Edge Computing Platform,” Applied Sciences, vol. 11, no. 8, pp. 1-16, 2021.
. Yu-Chen Lin, Chun-Liang Lin, Shih-Ting Huang and Cheng-Hsuan Kuo, “Implementation of an Autonomous Overtaking System Based on Time to Lane Crossing Estimation and Model Predictive Control,” Electronics, vol. 10, no. 18, pp. 1-22, 2021.
. Carlos Betancourt, Wen-Hui Chen, “Deep Reinforcement Learning for Port-folio Management on Markets with Dynamic Number of Assets,” Expert Sys-tems with Applications, vol. 164, pp. 1-10, 2021.
Ching Chun Tu
Department of Engineering and Technologies
National Science and Technology Council