CAST

Connected Autonomous Safe Technologies

Texas A&M Engineering Experiment Station
  • Home
  • About Us
  • People
  • Research
    • Active Research Projects
    • Archived Research Projects
  • Publications
  • Contact Us

Safe-Core Control Architecture for Autonomous Vehicles and ADAS Systems

Background

The “Architecture” of a (software) system represents the hierarchical decomposition of the system into its components and subsystems, the interfaces of the components and subsystems, and the connectivity between them. Using such an architecture as the underpinning for development of the controls and software system is “Architecture Based Development”.

An “Architecture Driven Development”approach to control system development is well-recognized as an important enabler for efficient control development processes. In particular, it allows for modular and distributed development, variant and configuration management, promotes reuse, and human-comprehensibility of complex (software) systems.

We believe that such an architecture based approach could also be leveraged to improve safety of ADAS systems and more broadly Autonomous and Connected Vehicles. This research intends to explore this hypothesis in detail.

Research Objectives, and anticipated Benefits

We would like to explore and develop a reference “Safe-Core Architecture” that becomes the basis for developing the AV Embedded Control Software. Some of the research questions addressed would include (but not be limited to):

  • Can we build a C/AV architecture to ensure Safe-Default actions. In particular, as we consider systems that have non-real-time (e.g. ROS based) and non-deterministic (e.g. ML based) components, can we provide default behavior that can guarantee some notion of safe behavior?
  • Can we model and analyze the architecture to have Provable or Verifiable Architectural Properties that promote Safety.
  • Can we build an Architecture that facilitates the build-up of “fault-tolerant” controls to address various faults, including sensor and actuator failures

Research Results

As part of the VAVS-TDP development, we already have a CAST Architecture for both system-level representation (that includes the physical plant and the controller) as well as a controller-level representation. This research will build upon this CAST Architecture.

Project Status

Just starting.

TDP Alignment:

We expect that this work will be cross-cutting and be aligned with all major TDPs

Research Themes Alignment:

 

Researchers:

  • Dr. Swami Gopalswamy
  • Dr. Srikanth Saripalli

External Links:

CAST
305 Doherty Bldg,
242 Spence St,
College Station, TX 77840

castinfo@tamu.edu

Texas A&M University College of Engineering
  • Accessibility
  • State Links and Policies
  • Privacy Policy
  • Website Feedback
  • Texas A&M Engineering Experiment Station

Copyright © 2023 · Texas A&M Engineering Experiment Station · All Rights Reserved