Review of “The Autonomous” Autonomous Systems: From Cars to the Moon

Related Vendors

FAULHABER Drive Systems

Taiwan Machine Tool & Accessory Builders' Association (TMBA)

SIMTOS (Seoul International Machine Tool Show)

The path to safe autonomous systems—which should also be as affordable as possible—is much more challenging than originally thought. To bring all stakeholders together, the company TTTech Auto—which, following its acquisition by NXP Semiconductors, now operates under the name “TrustMotion”—launched the “The Autonomous” initiative in 2017. At that time, the focus was clearly on autonomous driving, but now the initiative encompasses autonomous systems of all kinds.

At the Autonomous Main Event in Vienna this fall, the founding of a new association was therefore announced, operating under the name “The Autonomous – Global Platform for Safe Autonomous Systems and Robotics” (Figure 1). This independent and open multi-stakeholder organization aims to advance safe, secure, and robust autonomous systems through cross-industry collaboration. The credo of Ricky Hudi, Chairman of The Autonomous, plays a central role in this effort: Safety is a must—“Safety is non-negotiable,” as Hudi put it.

Founded by NXP, TrustMotion, Infineon, Qualcomm, and TTTECH, the association builds on the success of the Autonomous Initiative, which has brought together many stakeholders from industry, academia, politics, regulation, and research since its inception. “The future of autonomous systems depends on industry-wide collaboration,” explained Ricky Hudi, chairman of the Autonomous Association in Vienna. “By bringing together leading organizations from across the entire ecosystem, we are creating a neutral platform to accelerate innovation and lay the trustworthy foundations necessary for the widespread scaling of autonomous systems.” This platform is designed to promote collaboration in the automotive, robotics, and industrial automation sectors, as well as other application areas of autonomous systems.

Putting Physical AI into Practice

It was clear to all 600-plus attendees of the Autonomous Main Event that artificial intelligence (AI) will play a crucial role in autonomous systems, but “what really matters is putting Physical AI into practice,” explained Philipp von Schierstaedt, EVP & CSO of Consumer, Computing, and Communication at Infineon, in his presentation “The Future of Physical AI Is Based on Chips.”

He pointed out that the complexity of use cases in robotics is greater than in the automotive industry. While it is possible to draw on insights from automotive applications, the variety of use cases in robotics is much greater. For example, a car is controlled in two dimensions—by accelerating and braking, as well as by steering left and right. Robots, on the other hand, also sense and act in the vertical plane, and thus operate in a three-dimensional environment. While a vehicle’s autonomy is confined to the relatively narrow context of “driving,” (humanoid) robots, by their very nature, must perform or process many hundreds of different tasks, making their working environment considerably more complex than that of a vehicle. However, the fail-operational concepts and architectures developed in the automotive world also provide valuable services in robotics. It is therefore only a matter of time before the automotive industry receives valuable feedback from robotics. If the forecasts cited by Philipp von Schierstaedt prove accurate, the automotive sector should soon benefit from these insights from robotics: By 2030, 1.5 million humanoid robots are expected to be in use worldwide; by 2035, that number is projected to reach 11.6 million; and by 2050, a whopping 300 million.

Trust is the Foundation

“Practical deployment must be based on trust,” and to achieve this, autonomous systems require safety (functional safety) and security (cybersecurity), which must be deeply embedded across all hardware and software levels, von Schierstaedt noted. In this context, the ecosystem’s resilience acts as a gateway to scaling—and “for that, we need collaborative learning, virtualization, and supply chain resilience from the chip all the way to AI and the OEM level.”

“There is no market for physical AI without trust in the autonomous systems that surround us,” says Lars Reger, EVP & CTO of NXP, summing up the situation. “Trust begins with the right system architectures that combine safety, security, AI, and deterministic guardrails. With the Autonomous Association, we’re bringing industries together to create the safe and intelligent foundation necessary for the widespread scaling of autonomous systems.” In his keynote address, “How Robots Earn Their Place,” Péter Fankhauser, Co-Founder & CEO of ANYbotics, distilled trust in a human being into a mathematical formula, and then immediately presented the corresponding equivalent required for trust in an autonomous robot (Figure 2).

Autonomous Robots on the Moon

Julia Badger, Moon Base’s Surface Mobility Deputy Lead at NASA, gave a rather unusual presentation. She said that about 900 commands sent from Earth must be executed every day to keep the ISS operational. However, the ISS is only about 400 km (249 miles)  from Earth, while the Moon is, on average, 380,000 km (236,000 miles)  farther away. Remote control—an option that works well for cars and is just barely feasible for the ISS in certain exceptional cases—is therefore completely out of the question for operating autonomous systems on the Moon due to the long latency times, which effectively constitute control dead times.

On the Moon, therefore, full Level 5 functionality is required when Lunar Train Vehicles (LTVs) take on the “vital task of moving cargo from the landing zones to the actual mission sites.” The goal, therefore, is to transition in terms of autonomy from “humans in the loop” to “humans on the loop.” In addition to a robust design for critical components that operates in a fail-operational mode, she also emphasizes the importance of maintenance and inspection in terms of system design.

The Autonomous Systems Management (ASM) software is designed to ensure autonomous operation on-site—for example, on the Moon. This requires not only a distributed hierarchical architecture but also clear definitions of the interfaces and interdependencies.

For NASA, the top priority in closing the remaining technology gaps for the lunar mission is autonomous mobility on the lunar surface—including navigation. In second place is the unloading of the payload, along with its handling and processing by machines on the Moon. Only in third and fourth place do the systems for survival during the long night and shadow phases on the Moon, as well as resistance to the fine but sharp-edged lunar dust, rank. This clearly demonstrates just how important autonomous mobility activities are to NASA.