ANYbotics is a Swiss-based developer of AI-driven robot inspection solutions and was set up to tackle the challenges around delivering safety and efficiency in some of the world’s toughest industrial environments.
The company’s ANYmal robots are intended for autonomous operation and can deliver advanced mobility and real-time data collection and are suitable for remote operations and predictive maintenance.
First conceived back in 2009 the company has received over $130m in funding from European and Silicon Valley investors and has offices in both Zurich and San Francisco.
Its robots have been deployed around the world and across the energy, power, metals, mining, and chemical industries, where they are used to augment industrial workforces and to address labour shortages.
“ANYbotics was founded as a spin-off from ETH Zurich, and we saw our first sales back in 2017,” explained Péter Fankhauser, the company’s co-founder and CEO.
“My doctorate at ETH Zurich involved the development of perception and motion planning algorithms for legged robots that could be deployed in challenging environments. Working with colleagues our first prototype was built back in 2009. Our aim was to transform the way industrial facilities are operated.”
According to Fankhauser the decision to go with the four-legged concept was inspired by watching mountain goats. “We examined a number of form factors, but it turned out to be ideal in terms of stability.”
“We want to empower businesses with cutting-edge robotics technology, so that robots and humans can collaborate seamlessly.
“These businesses need access to data but they don’t have enough people, and those they have are obliged to work in dangerous situations and it can be very repetitive. When it comes to monitoring facilities it’s not possible to place sensors everywhere and it can be extremely costly. You could use drones but they’re expensive and flight times are limited.
“So, our focus was robotics, and our first prototype could both walk and self-right. Our first ANYmal robots appeared in 2015 and had modular actuators and lidar-based navigation.”
The biggest issue with these early prototypes was that it was difficult to get the necessary components, according to Fankhauser.
“Our four-legged robots had a broad footprint for maximum stability and they needed to be lightweight – but that raised the question as to where could we get the components we needed? That was a real challenge. When we first started, we were having to develop our own motors and sensors but today robotics technology has advanced dramatically.
“In truth, the only thing holding us back are the limits of human ingenuity. Artificial intelligence has enabled our robots to be able to react and adapt in real time.
"The issue is less about the technology and more about educating customers in how to use our robots. They need to understand automation and you need to take them on a journey.”
To that end the company employs field engineers to work closely with customers providing on the job training which enables customers to create their own mission, make sense of the data that’s collected and then to develop and plan new missions.
ANYmal uses Pharos, a 3D software for simultaneous localisation and mapping (SLAM), to navigate complex industrial facilities.
“It provides users with centimetre-accurate mapping through sensor fusion of lidar and depth cameras, providing area-wide mapping with a range of 4 km in a single pass,” explained Fankhauser. “An employee can use a tablet to remotely control the robot, to tell it what to do, and to set up a specific mission. The data that’s collected can then be edited, transferred back to the robot, and a mission set up in a matter of a couple of days.”
Above: An ANYmal inspection robot in operation
Autonomous operation
The company’s latest platform, the ANYmal X can carry out its missions autonomously or remotely and is equipped with a high-resolution camera, laser imaging, gas sensor technology, thermal cameras and wide-spectrum directional microphones.
These components allow for mapping of an installation in images and 3D point clouds. The platform can check for anomalies such as temperature, leak detection or simply read instruments.
Images are transmitted via wireless communication (WiFi or 4G LTE) to databases and can then be used to ensure that any interventions are carried out where they are really needed, although in some of the extreme environments in which these robots are deployed – where metal and concrete, for example, are used extensively – wireless communications wont be possible, so data is stored to be downloaded later.
“The use of our technology is about managing assets more effectively. Asset failure can trigger substantial downtime and expense, which can impact both safety and productivity,” said Fankhauser. “Our robots can conduct hundreds of visual, thermal, and acoustic inspections on a daily basis, and now we have developed a new asset management platform – Data Navigator – that can transform the data that’s collected into actionable insights.
“We can provide a comprehensive view of asset health, enabling preventive maintenance strategies that minimise downtime and optimise operations.”
According to Fankhauser Data Navigator was a response to customer requests.
“As we deployed more robots, so we get much more feedback. What works, how to enhance software capabilities or the types of robots that we should be looking to deploy. With AI, autonomy and the data layer everything is open ended. We’re dealing with many large companies with hundreds of different facilities in which we can deploy our robots.”
Asset integrity is critical for the industries ANYBotics operates in and Data Navigator can be deployed to centralise multiple key asset condition data types, such as thermal, acoustic, visual, and gas readings.
“It’s about simplifying asset health assessments and provides trend monitoring for preventive maintenance. By using Data Navigator raw sensor outputs can be transformed into actionable intelligence as it’s collected by our ANYmal robots. Data Navigator visualises multi-modal insights giving maintenance teams a clear picture of asset health, helping them spot performance issues before they escalate to failures,” Fankhauser explained.
"Robotic inspection offers maximum value when the resulting data is readily available and actionable," said Fankhauser, “and Data Navigator is the key to unlocking that value. It can be used to monitor trends and make informed decisions, optimise maintenance schedules, and ultimately ensure continued plant uptime and the efficiency of operations."
“Sensitive data is kept secure, and the platform’s intuitive interface requires no specialised robotics expertise, ensuring broad accessibility across teams. The solution is ready to use out-of-the-box and can integrate with enterprise systems like SAP and IBM when required.
“Effective industrial asset management depends on making advanced analytics accessible to everyone without requiring specialised training in either robotics or data science,” according to Fankhauser. “We believe that Data Navigator empowers teams to make informed decisions quickly and confidently.”
The ability to transform inspection data into actionable insights is becoming increasingly important in terms of managing assets and the work being conducted by ANYbotics is helping to lay the groundwork for the next generation of intelligent industrial maintenance.
“By combining autonomous robotics with intelligent analytics, industries can move beyond reactive maintenance to genuinely preventative maintenance strategies, ensuring operational continuity and maximising returns on critical infrastructure investments,” concluded Fankhauser.
