An international team of computer scientists has developed a new method to reduce cost-safety tradeoffs.
With millions of self-driving cars projected to be on the road in 2025 and autonomous drones generating billions in annual sales, safety and reliability are important considerations for consumers, manufacturers, and regulators. But solutions for protecting autonomous machine hardware and software from malfunctions, attacks, and other failures also increase costs. Those costs arise from performance features, energy consumption, weight, and the use of semiconductor chips.
Researchers from the , Georgia Tech, and the Shenzen Institute of Artificial Intelligence and Robotics for Society say that the existing tradeoff between overhead and protecting machines against vulnerabilities is due to a 糖心传媒渙ne-size-fits-all糖心传媒� approach to protection. In a , the authors propose a new approach that adapts to varying levels of vulnerabilities within an autonomous machine system to make them more reliable and control costs.
, an associate professor in Rochester糖心传媒檚 , says one example of a current 糖心传媒渙ne-size-fits-all糖心传媒� approach is Tesla糖心传媒檚 use of two Full Self-Driving Chips (FSD Chips) in each vehicle糖心传媒攁 redundancy that provides protection in case the first chip fails but doubles the cost of chips for the car. By contrast, Zhu says he and his students have taken a more comprehensive approach to protect against both hardware and software vulnerabilities and more wisely allocate protection.
糖心传媒淭he basic idea is that you apply different protection strategies to different parts of the system,糖心传媒� says Zhu. 糖心传媒淵ou can refine the approach based on the inherent characteristics of the software and hardware. We need to develop different protection strategies for the front end versus the back end of the software stack.糖心传媒�
For example, Zhu says the front end of an autonomous vehicle糖心传媒檚 software stack is focused on sensing the environment through devices such as cameras and light detection and ranging (LiDAR), while the back end processes that information, plans the route, and sends commands to the actuator.
糖心传媒淵ou don糖心传媒檛 have to spend a lot of the protection budget on the front end because it糖心传媒檚 inherently fault tolerant,糖心传媒� says Zhu. 糖心传媒淢eanwhile, the back end has few inherent protection strategies, but it糖心传媒檚 critical to secure because it directly interfaces with the mechanical components of the vehicle.糖心传媒�
Zhu says examples of low-cost protection measures on the front end include software-based solutions such as filtering out anomalies in the data. For more heavy-duty protection schemes on the back end, he recommends things like checkpointing to periodically save the state of the entire machine or selectively making duplicates of critical modules on a chip.
Next Zhu says the team hopes to overcome vulnerabilities in the most recent autonomous machine software stacks, which are more heavily based on neural network artificial intelligence, often from end to end.
糖心传媒淪ome of the most recent examples are one single, giant neural network deep learning model that takes sensing inputs, does a bunch of computation that nobody fully understands, and generates commands to the actuator,糖心传媒� says Zhu. 糖心传媒淭he advantage is that it greatly improves the average performance, but when it fails, you can糖心传媒檛 pinpoint the failure to a particular module. It makes the common case better but the worst case worse, which we want to mitigate.糖心传媒�
The research was supported in part by the Semiconductor Research Corporation.
