The Rise of Safety-Certified AI: How MIPS and Green Hills are Paving the Way
The convergence of artificial intelligence (AI) and safety-critical systems is rapidly accelerating, particularly in automotive and industrial applications. A recent collaboration between MIPS, a GlobalFoundries company and Green Hills Software aims to streamline the development and certification of these complex systems, promising faster time-to-market and increased reliability.
The Necessitate for Safety in AI-Powered Systems
As AI moves beyond simple automation and into applications controlling physical processes – like electric vehicle motor control or industrial robotics – the need for robust safety measures becomes paramount. Traditional software development approaches often fall short when dealing with the unpredictable nature of AI algorithms. Functional safety standards, such as ASIL-D (Automotive) and SIL 3/4 (Industrial), are designed to mitigate risks and ensure systems operate predictably even in the face of failures.
The challenge lies in achieving certification to these standards while still harnessing the power of AI. This requires a complete hardware-software pathway designed from the ground up with safety in mind.
MIPS and Green Hills: A Synergistic Approach
MIPS’ M8500 RISC-V microcontroller architecture, optimized for real-time, multi-threaded compute, forms the foundation of this new approach. RISC-V’s open nature allows for customization and integration of safety features. Paired with Green Hills Software’s certified tools – including compilers, a real-time operating system (RTOS) called µ-velOSity, and static analysis tools – developers gain a comprehensive suite for building and verifying safety-critical applications.
The jointly developed Safety SDK will integrate these components, providing a complete development path to safety-certified production. Early access customers are already working with the platform, focusing on motor control demonstrations and evaluation platforms.
Real-Time Performance and Deterministic Behavior
A key benefit of this collaboration is the focus on real-time performance and deterministic behavior. In safety-critical applications, it’s not enough for a system to simply *eventually* respond to a stimulus; it must respond *within a guaranteed timeframe*. The M8500 architecture, combined with the µ-velOSity RTOS, is designed to deliver this level of predictability.
This represents particularly significant in applications like industrial robotics, where a delayed response could lead to collisions or damage. Similarly, in electric vehicles, precise and timely control of the motor and power electronics is essential for safety and efficiency.
The Role of Virtual Platforms and RISC-V
MIPS leverages virtual platforms to enable a modular, standards-based approach to development. This allows developers to test and refine their designs in a simulated environment before committing to hardware, reducing development time and costs. The use of the open RISC-V instruction set architecture further enhances flexibility and customization.
Future Trends: Physical AI and Edge Computing
This collaboration points to a broader trend: the rise of “Physical AI.” This refers to AI algorithms that directly control physical systems, rather than simply analyzing data. Edge computing, where data processing occurs closer to the source, is also crucial. Processing data locally reduces latency and improves security, both essential for safety-critical applications.
As IoT devices develop into more prevalent in transportation, industrial, and medical markets, the demand for safe, efficient, and future-proof solutions will only increase. The combination of advanced processor architectures, certified software tools, and edge computing will be key to meeting this demand.
FAQ
Q: What are ASIL-D and SIL 3/4?
A: These are functional safety standards used in automotive (ASIL-D) and industrial (SIL 3/4) applications to define levels of risk and required safety measures.
Q: What is RISC-V?
A: RISC-V is an open-source instruction set architecture (ISA) that allows for greater flexibility and customization compared to proprietary ISAs.
Q: What is an RTOS?
A: A Real-Time Operating System (RTOS) is designed to provide predictable and timely responses to events, crucial for safety-critical applications.
Q: What is the benefit of using a virtual platform?
A: Virtual platforms allow developers to test and refine their designs in a simulated environment, reducing development time and costs.
Did you know? Green Hills Software’s INTEGRITY-178 RTOS was the first operating system to be certified to EAL 6+, High Robustness, the highest level of security ever achieved for any software product.
Pro Tip: When developing safety-critical systems, prioritize early and continuous verification throughout the entire development lifecycle.
Learn more about the latest advancements in safety-certified AI and explore how these technologies are shaping the future of automotive and industrial automation. Share your thoughts and experiences in the comments below!
