A matched case-control analysis compared autonomous vehicle (AV) and human-driven vehicle (HDV) accidents, using 2100 AV and 35,113 HDV accident data. The study found that AVs generally have a lower accident risk than HDVs in most scenarios, but higher risks during dawn/dusk and turning conditions, with 5.25 and 1.98 times higher accident rates, respectively. AVs equipped with Advanced Driving Systems (ADS) showed reduced accident likelihood in most conditions, but increased risk in complex scenarios like turning. The analysis revealed that AVs are safer in rear-end and broadside accidents compared to HDVs, but more prone to head-on collisions. AVs also had higher accident rates in rainy conditions and during dawn/dusk. The study used a matched case-control logistic regression model to assess accident risk factors, finding that AVs had lower odds of accidents in rainy weather and during dawn/dusk. However, AVs had higher accident rates in turning conditions. The study also found that AVs were less likely to be involved in accidents when proceeding straight, but more likely in run-off road and entering traffic lane conditions. The research highlights the need for improved sensors, algorithms, and design to enhance AV safety in challenging conditions. The study used real-world accident data and multi-source information to analyze AV and HDV accident factors, providing insights into accident risk differences between AVs and HDVs. The findings suggest that AVs can reduce accident risks in many scenarios but face challenges in complex driving conditions. The study also identified key factors influencing accident risk, including weather, road conditions, and pre-accident movements. The research contributes to understanding AV safety and highlights the importance of addressing safety challenges in AV development. The study's results emphasize the need for continued research and improvements in AV technology to enhance safety and reduce accident risks.A matched case-control analysis compared autonomous vehicle (AV) and human-driven vehicle (HDV) accidents, using 2100 AV and 35,113 HDV accident data. The study found that AVs generally have a lower accident risk than HDVs in most scenarios, but higher risks during dawn/dusk and turning conditions, with 5.25 and 1.98 times higher accident rates, respectively. AVs equipped with Advanced Driving Systems (ADS) showed reduced accident likelihood in most conditions, but increased risk in complex scenarios like turning. The analysis revealed that AVs are safer in rear-end and broadside accidents compared to HDVs, but more prone to head-on collisions. AVs also had higher accident rates in rainy conditions and during dawn/dusk. The study used a matched case-control logistic regression model to assess accident risk factors, finding that AVs had lower odds of accidents in rainy weather and during dawn/dusk. However, AVs had higher accident rates in turning conditions. The study also found that AVs were less likely to be involved in accidents when proceeding straight, but more likely in run-off road and entering traffic lane conditions. The research highlights the need for improved sensors, algorithms, and design to enhance AV safety in challenging conditions. The study used real-world accident data and multi-source information to analyze AV and HDV accident factors, providing insights into accident risk differences between AVs and HDVs. The findings suggest that AVs can reduce accident risks in many scenarios but face challenges in complex driving conditions. The study also identified key factors influencing accident risk, including weather, road conditions, and pre-accident movements. The research contributes to understanding AV safety and highlights the importance of addressing safety challenges in AV development. The study's results emphasize the need for continued research and improvements in AV technology to enhance safety and reduce accident risks.