Sensor Data Confidence Scoring for Secure Automated Machines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated machines, such as self-driving vehicles, are vulnerable to security breaches when sensors are compromised by malware, leading to potential safety risks due to erroneous driving commands.
Innovation Solution
A communication system that includes a security server and a sensor security evaluation application, which assigns a security confidence score to sensor data based on vulnerability, location, and staleness factors, allowing the automated machine to differentiate and mitigate risky data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If sensor data is processed without security verification, then processing speed is improved, but security reliability deteriorates
Solution Approach 1:
The system performs security verification in advance by calculating security confidence scores for sensor data before the automated machine processes it. This preliminary security assessment prevents malicious data from entering the processing pipeline, maintaining both security reliability and processing speed by filtering threats beforehand rather than during processing.
2Reliability
If security verification is performed on all sensor data, then security reliability is improved, but processing time increases
Solution Approach 1:
The system applies security verification selectively rather than uniformly to all sensor data. By calculating security confidence scores and comparing them against thresholds, the system focuses verification efforts on data with potential security risks while allowing clearly safe data to pass through with minimal processing, thus maintaining security reliability without excessive time loss.
3Reliability
If security confidence scoring is implemented, then security differentiation is improved, but system complexity increases
Solution Approach 1:
The system introduces a security server as an intermediary component that handles the complex security confidence score calculation and verification processes. This separates the security verification function from the main automated machine processing system, improving security differentiation while managing system complexity by isolating the complex security logic in a dedicated external service.
Data Source
Figure 1
Figure 2~3
AI summary
Systems, methods, and software can be used to provide secure sensor data. In some aspects, a computer-implemented method includes: receiving, at a sensor security evaluation application executing on a device, sensor data from a sensor on the device; determining, by the sensor security evaluation application, a security confidence score associated with the sensor data; and transmitting, from the sensor security evaluation application, the security confidence score and the sensor data to a smart machine processor on the device.