特斯拉案例研究-新能源电动汽车公司英文版.pdf

Case Study: Tesla Motors Continue Creating Self-driving cars or not? Sayidana Brannan-Douglas | Hamza Mahmood | Bowen Zheng High Level Overview of the Problem ● Increased humans trust and reliance on new systems. Main Sources of concern, % of respondents* ● Adaptive cruise control often fails to stop in the case of stationary objects. ● 3 out of 6 fatal accidents linked to autopilot feature were associated with Tesla’scar. ● Tesla’s in-house technology is not the safest compared to its competitors * Based on responses by car buyers in China, Germany, Japan and S. Korea (n = 3500) Sub-Problems of Problems To Address ● Safety ● Socio-Ethical ● Financial Safety Challenges Lack of LiDAR system Flaw in Sensor Layout Sub-optimal Performance of Car Computer Safety Lack of Sensor: LiDAR ● Light Detection and Ranging (LiDAR) scanning is the state-of-the-art in surveying technology. ● More advanced than the sonar and radar. ● Instead of using sound or radio waves to scan its environment, LiDAR utilizes laser light pulses. Safety Lack of Sensor: LiDAR ● Competing companies are using LiDAR, even LiDAR matrix for more accurate mapping and ranging. Except Tesla… ● Major Disadvantages: ○ Super expensive ($75k per unit) ○ Redesign and re-write the program of Autopilot software. Safety Flaw of sensors: Layout Increase the potential possibility of crash due to Blind zones not detect or detect too late. Safety Short of Car computer: under-performance ● Currently using: ○ Nvidia Drive PX2 Platform ○ 8000 Gflops ○ 20