Vehicle Controller Sandbox Partition for Safe Software Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern vehicles with increasingly complex software-driven systems face challenges in verifying functional components under real conditions, as development vehicles often lack necessary subsystems and assumptions about software behavior prove suboptimal, leading to complex regression tests and difficulties in continuously improving software functions during both development and serial operation.
Innovation Solution
A method utilizing a vehicle controller with two partitions, where a regular application operates on one partition for normal vehicle functions and a sandbox on another for testing new or modified functions, allowing continuous learning and iterative development without interfering with the primary system, using read-only access to prevent write interventions and enabling data collection for verification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a regular application is executed on a vehicle controller for normal operation, then the vehicle functions are maintained, but the verification of new software functions under real conditions becomes difficult
Solution Approach 1:
The vehicle controller is divided into two separate partitions: a first partition for executing the regular application that maintains normal vehicle functions, and a second partition for providing a sandbox environment to test new software functions. This segmentation allows both reliable operation and software verification to occur simultaneously without interfering with each other.
2Productivity
If new software functions are tested on the vehicle controller, then continuous improvement is enabled, but the system may be changed, disturbed or damaged
Solution Approach 1:
The testing environment is extracted as a separate sandbox on the second partition of the vehicle controller. This sandbox provides read-only access to data from the first partition but prevents any write interventions that could modify or damage the regular application. New software functions can be tested aggressively without risking the stability of the production system.
Solution Approach 2:
The sandbox environment is prepared in advance as a protective layer that absorbs potential errors or disruptions from testing new software. The read-only access restriction acts as a cushion that prevents testing activities from propagating harmful effects to the regular application, allowing rapid development while maintaining system reliability.
3Measurement precision
If development vehicles are equipped with all necessary subsystems for complete testing, then verification accuracy improves, but device complexity and cost increase
Solution Approach 1:
The vehicle controller itself is made multi-functional by incorporating both the regular application and the sandbox environment within a single device. This allows the controller to serve dual purposes: maintaining normal vehicle functions while simultaneously providing a testing environment for new software, eliminating the need for separate development vehicles equipped with additional subsystems.
Solution Approach 2:
The sandbox creates a virtual copy of the operational environment on the second partition, allowing new software to be tested against real vehicle data without requiring physical duplication of all vehicle subsystems. The read-only access to the first partition provides a realistic operational context for testing while avoiding the complexity of equipping development vehicles with complete subsystems.
4Loss of time
If assumptions about software behavior are made at the beginning of series production, then development time is reduced, but the assumptions are later found to be not optimal or not applicable
Solution Approach 1:
The sandbox environment enables continuous testing and verification of software behavior assumptions throughout the production lifecycle. Rather than making static assumptions at the beginning of series production, the read-only access to operational data allows continuous validation and refinement of software behavior models, ensuring assumptions remain accurate while maintaining rapid development pace.
Data Source
AI summary
A method for operating a vehicle controller. A regular application for normal operation of the vehicle controller is executed on a first partition. A sandbox is provided on a second partition. In the sandbox, an application to be tested is executed. The sandbox is provided such that read access to the regular application is possible therefrom. Signals and/data used in the regular application during normal operation are obtained and processed by the application to be tested. A corresponding vehicle controller is also provided.