Improve trust in autonomous technology

Trust and assurance — from consumers, the public, and governments — will be critical issues for AI and the autonomous tech space in the coming year. Yet earning that trust will require fundamental innovations in how autonomous systems are tested and evaluated, according to Shawn Kimmel, executive director of EY-Parthenon Quantitative Strategies and Solutions at Ernst & Young LLP. Fortunately, the industry now has access to innovative techniques and emerging methods that promise to transform the field.

The new autonomy environment

Automation has always been touted as a replacement for “boring, dirty and dangerous” jobs, and this continues to be the case, whether working in underground mines, maintaining offshore infrastructure or, because of the pandemic, in medical facilities. Protecting people from danger in sectors as essential and varied as energy, raw materials and health remains a laudable objective.

But self-managed technologies are now moving beyond those applications, finding ways to improve efficiency and convenience in everyday spaces and environments, Kimmel says, through innovations in computer vision, artificial intelligence, robotics , materials and data. Warehouse robotics have evolved from glorified streetcars shuttling materials from A to B to intelligent systems that can move freely through space, identify obstacles, change routes based on stock levels, and manage delicate items. In surgical clinics, robots excel in microsurgical procedures in which the slightest human tremor has negative impacts. Startups in the autonomous vehicle sector are developing applications and services in niches such as mapping, data management and sensors. Robot taxis are already in commercial operation in San Francisco and are expanding from Los Angeles to Chongqing.

As autonomous technology enters more and more settings, from public roads to medical clinics, safety and reliability become both more important to prove and more difficult to ensure. Autonomous vehicles and unmanned aerial systems have already been implicated in accidents and casualties. “Mixed” environments, including both human and autonomous agents, have been identified as posing new security challenges.

The expansion of autonomous technology into new areas is driving a growing number of players, from equipment manufacturers to software startups. This “system of systems” environment complicates testing, security, and validation standards. Longer supply chains, as well as more data and connectivity, introduce or increase security and cyber risk.

As the behavior of autonomous systems becomes more complex and the number of stakeholders increases, security models with a common framework and terminology and interoperable testing become necessities. “Traditional systems engineering techniques have been pushed to their limits when it comes to stand-alone systems,” says Kimmel. “There is a need to test a much broader set of requirements, as autonomous systems perform more complex tasks and safety-critical functions.” This need in turn generates interest in the search for efficiencies, in order to avoid soaring test costs.

This requires innovations like predictive safety performance metrics and preparedness for unexpected “black swan” events, Kimmel says, rather than relying on conventional metrics like mean time between failures. It also requires ways to identify the most valuable and impactful test cases. Industry needs to increase the sophistication of its testing techniques without making the process unduly complex, expensive or inefficient. To achieve this goal, it may be necessary to manage the set of unknowns in the operating mandate of autonomous systems, reducing the test and security “state space” from a semi-infinite state to a testable set of conditions.

Test, test

The toolkit for security, testing, and assurance of autonomous systems continues to evolve. Digital twins have become a development asset in the field of autonomous vehicles. Virtual and hybrid “loop” test environments enable system-of-system testing that includes components developed by multiple organizations across the supply chain, and reduce the cost and complexity of real-world testing through to digital augmentation.