In recent years, underground settings are becoming increasingly relevant to global security and safety. Factors such as rising populations and urbanization are requiring military and civilian first responders to perform their duties in below-ground conditions, in human-made tunnels, underground urban spaces, and natural cave networks. Military and Security forces also have interest in developing robots to target tunnels dug by Militaries and terrorists . Criminals, terrorists and military have been using tunnels since long time to evade detection as surface based detection methods are ineffective underground .The tunnel threat is also a serious and growing concern to U.S and Mexico, as they enable human trafficking and smuggling of drugs and weapons across the border.
DARPA first announced its Subterranean (SubT) Challenge in 2018 with aim to explore new approaches to rapidly map, navigate, and search underground environments. Teams from around the world were invited to propose novel methods for tackling time-critical scenarios through unknown courses in mapping subsurface networks and unpredictable conditions, which are too hazardous for human first responders. The underground environments DARPA wanted to test in the competition included human-made ones, like subway systems, sewers, and mines, and natural ones, like caves and tunnels.
“One of the main limitations facing warfighters and emergency responders in subterranean environments is a lack of situational awareness; we often don’t know what lies beneath us,” said Timothy Chung, program manager in DARPA’s Tactical Technology Office (TTO). “Building an accurate three-dimensional picture is a key enabler to rapidly and remotely exploring and searching subterranean spaces.”
The US Defense Advanced Research Projects Agency (DARPA) developed advanced technologies and methodologies for 3D mapping and surveying of underground environments under the agency’s Subterranean (SubT) Challenge, which aims to equip troops and first responders with superior capabilities to effectively execute missions underground. The agency explored the use of mobile robots for search and rescue to avoid the dangers miners and explorers. “We’ve reached a crucial point where advances in robotics, autonomy, and even biological systems could permit us to explore and exploit underground environments that are too dangerous for humans,” said TTO Director Fred Kennedy.
The second phase of the DARPA Subterranean Challenge concluded in Feb 2020 with the completion of the Urban Circuit. Similar to the challenge’s first phase, the urban circuit challenged teams to use robots and drones to rapidly map, navigate and search underground environments. Taking first in the competition was CoSTAR, a 12-robot, 60-person team led by NASA’s Jet Propulsion Laboratory (there were also winners declared for a separate, virtual competition).
Two previous scored events – Tunnel and Urban Circuits – featured both Virtual and Systems Competitions. DARPA has made the difficult decision to proceed only with the Virtual Competition for the Cave Circuit, due to safety considerations surrounding COVID-19.
The Final Event, planned for the latter half of 2021, will combine elements from each of the three subdomains. As with the previous two scored circuit events, the Systems and Virtual Competitions will take place in parallel with DARPA-funded and self-funded teams competing side-by-side. Teams in the Systems Competition will compete for up to $2 million in the Final Event Systems Competition. Teams in the Virtual Competition will compete for up to $1.5 million in the Virtual Final Event Virtual Competition.
Tunnel detection and mapping is becoming a critical capability for the U.S. military as well as the Department of Homeland Security. According to the Science and Technology directorate at DHS, finding illicit tunnels and underground passages along the Southern border is largely based on “random tips and laborious human intelligence” and not on detection technology.
Subterranean environment Challenges
DARPA Tactical Technology Office (TTO) programme manager Timothy Chung said: “What makes subterranean areas challenging for precision mapping and surveying, such as lack of GPS, constrained passages, dark or dust-filled air, is similar to what inhibits safe and speedy underground operations for our warfighters. “Instead of avoiding caves and tunnels, we can use surrogates to map and assess their suitability for use. Through the DARPA Subterranean Challenge, we are inviting the scientific and engineering communities—as well as the public—to use their creativity and resourcefulness to come up with new technologies and concepts to make the inaccessible accessible.”
The challenges presented by subterranean environments can vary drastically across subdomains, which can include human-made tunnel systems, urban and municipal underground infrastructure, and natural cave networks. Tunnels can extend many kilometers in length and can include highly constrained passages, multiple levels, and vertical shafts. Alternatively, urban underground environments are often more structured and constructed out of man-made materials, but can have complex layouts that cover multiple stories and/or span multiple city blocks. Natural cave networks often have irregular geological structures, with both constrained passages and large caverns, and unpredictable topologies often stretching large distances in extent and depth.
Even under ideal conditions, these complex environments present significant challenges for subterranean situational awareness. However, in time-sensitive scenarios, whether in active combat operations or disaster response settings, warfighters and first responders alike are faced with a range of increased technical challenges, including difficult and dynamic terrains, unstable structures and obstacles, degraded environmental conditions, severe communication constraints, and expansive areas of operation. In many cases, these environments pose too great a risk to deploy personnel, and current technologies fail to provide the necessary rapid and remote mapping, navigation, and search capabilities.
The multi-faceted nature of these problems presents both a need and an opportunity for breakthrough innovations for public safety scenarios as well as a wide range of military, academic, and commercial applications, including infrastructure inspection, oil/gas/mining, construction, archeology, and scientific exploration. The DARPA Subterranean Challenge (a.k.a. the SubT Challenge) aims to bring together multi-disciplinary teams and industries across disparate fields to establish a broader research community and develop these innovative leap-ahead capabilities.
Subterranean Challenge: Program Goals
The primary goal of the DARPA Subterranean Challenge is to discover innovative solutions that can rapidly and remotely map, navigate, and search complex environments, including human made tunnel systems, urban and municipal underground infrastructure, and natural cave networks.
Given the complex problem space presented by these environments, the DARPA Subterranean Challenge seeks to inspire and realize technological breakthroughs to offer key insights into:
Disruptive concepts of operations that both enable and exploit the capability to conduct rapid and autonomous subterranean missions; and
Composition of system capabilities to offer freedom of mobility at operationally relevant speeds in complex, unpredictable, and diverse subterranean environments.
Technologies that are designed to provide accurate and high-resolution precise and reproducible survey points with no dependence on substantial infrastructure are of particular interest. Furthermore, technologies should be able to support easy manipulation, interpreted, and rendered data products into 3D mesh objects.
The variety of built-in challenge elements and the competition structure itself are intended to address the secondary goal of increasing the diversity, versatility, cost-effectiveness, and robustness of relevant technologies and systems capable of addressing the myriad needs of a wide range of environments rather than serving single-purpose uses with specifically tailored solutions.
Another goal of the DARPA Subterranean Challenge is to establish a collaborative community by bringing together multi-disciplinary teams and cross-cutting approaches across disparate fields to tackle the autonomy, perception, networking, and mobility needs of the subterranean domain. To encourage a broader range of participants, the SubT Challenge includes both a physical Systems competition as well as a software-only Virtual competition. Teams in the Systems competition will develop and demonstrate physical systems in real-world environments, which will emphasize the interdisciplinary nature of fielding integrated solutions. Teams in the Virtual competition will use virtual models of systems, environments, and terrain to compete in simulation-based events that will focus effort on software-driven innovations.
To initiate such cross-cutting approaches, DARPA will develop the SubT Virtual Testbed comprising an extensible and validated simulation environment, automated testing and assessment tools, and associated software support infrastructure to be provided as government-furnished equipment (GFE). Teams in both the Systems and Virtual competitions will leverage this suite of GFE simulation tools to accelerate the development and evaluation of their proposed solutions.
SubT is divided into four circuits spread over three years. With each, teams program their robots to navigate a complex underground course. Like previous challenges, DARPA-funded and self-funded teams will compete side by side. The competition will consist of two tracks. Participants on the systems track will need to develop technology like a tracked vehicle or walking robot that can traverse a physical test course. Teams on the software track will have to present their work in a simulated environment. Teams in both tracks will compete in three preliminary Circuit events and the Final event. Each Circuit event will explore the difficulties of operating in a specific underground environment. The first will focus on human-made tunnel systems. The second will focus on underground urban environments such as mass transit and municipal infrastructure. The third will focus on naturally occurring cave networks. They will have to demonstrate their technology in each of those three subdomains – first one at a time, and then in all three at once – earning scores based on speed, maneuverability, the accuracy of the maps they produce, and their ability to identify objects.
The first contest, held in August 2019 , took place in a mine. For the most recent, called the Urban Circuit, teams raced against one another in an unfinished power plant in Elma, Washington. Each team’s robots searched for a set of 20 predetermined objects, earning a point for each find. For the Urban Circuit, CoSTAR earned 16 points; the No. 2 team, with 11 points, was Explorer, led by Carnegie Mellon University.
The Final event, planned for 2021, will incorporate diverse challenges from various underground environments.The varied nature of the subdomains DARPA has designed the competition around means any proposal will have to maneuver through tight spaces, rugged terrain, man-made impediments, and natural obstacles – sometimes all on the same missions. The winner of the Systems track will take home a $2 million prize, while the winner of the Virtual track will earn $750,000. The deployment of innovative, enhanced technologies could accelerate the development of critical lifesaving capabilities.
The SubT Challenge physical Systems and software-focused Virtual competitions aim to create a community of multi-disciplinary teams from distinct fields to foster breakthrough technologies in autonomy, perception, networking, and mobility for underground environments.
Teams in the Systems competition will develop and demonstrate physical systems in real-world environments. DARPA has selected seven teams to compete in the funded track of the Systems competition: Carnegie Mellon University; Commonwealth Scientific and Industrial Research Organisation, Australia; iRobot Defense Holdings, Inc. dba Endeavor Robotics; Jet Propulsion Laboratory, California Institute of Technology; University of Colorado, Boulder; University of Nevada, Reno; and University of Pennsylvania
Teams in the Virtual competition will use simulation models and physics-based environments focusing on software-driven advances. The following organizations have received a contract to compete in the DARPA-funded track of the Virtual competition: Michigan Technological University and Scientific Systems Company, Inc.
Teams participating in the challenge will be tasked with designing and developing novel solutions that address the challenges of subterranean environments in circuits for each of the subdomains to include tunnel systems, urban underground, and natural cave networks, culminating with a final event encompassing elements from all three environments.
Virginia-based firm Scientific Systems Company was awarded a contract worth up to nearly $500,000 in July 2019. On August 8, Michigan Technological University got a contract worth up to $750,000, and Virginia-based company iRobot Defense Holdings received an award of up to $4.46 million. Open Robotics, which oversees development of the open-source Robotic Operating System (ROS), was a key player in the DRC. Many of the DRC competitors used platforms running ROS. Open Robotics also offers Gazebo, a physics engine and interface that allows developers to virtually simulate robots in real-world scenarios. Open Robotics hosted a virtual stage of the DRC, and it will run a similar virtual stage of the new SubT challenge.
“Open Robotics is excited to be working with DARPA on the virtual track of the SubT,” says Brian Gerkey, CEO of Open Robotics. “In this way, teams will be able to test, train and ultimately compete wholly within a simulated environment. These teams are going to be pushing the envelope, so being able to simulate robot behavior is way more efficient than constantly building and re-building hardware.”
Feasible robotics solutions
Team BARCS (Bayesian Adaptive Robot Control System) includes members of the Michigan Technological University and Michigan Tech Research Institute. The team views the challenge as “a problem in multi-agent coordination in highly resource-constrained settings. Resources in this case include agent lifespan, sensing ability, communications connectivity, among others. Our solution is inspired by the need to optimize the joint capabilities of the team as well as the utilization of their resources. We are leveraging the mathematical strengths of our team to develop principled, generalizable, and novel solution strategies.”
Flyability has teamed with the Autonomous Robots Lab at the University of Nevada, the Robotic Systems Lab of ETH Zurich, the Autonomous System Lab of ETH Zurich, and the HiPeR Lab of U.C. Berkeley, and Sierra Nevada Corporation for the challenge. The association is called CERBERUS. For subterranean deployments team CERBERUS visualizes the robotic solution through the collaboration of walking and flying robots. With the combination of best products and research projects, it will provide field experts with an autonomous, robust, and reliable way to fulfil their mission even in an unpredictable and adverse situation.
“The environments where the challenge is taking place have a lot in common with those where our customers are deploying Elios daily. Taking part in the prestigious Subterranean DARPA challenge is an opportunity to collaborate with university laboratories and companies which are the best in their R&D fields.” Stated Adrien Briod, CTO of Flyability. “Successfully completing these missions will require multiple robots, including both drones and ground vehicles,” said in a press release Sebastian Scherer, who will lead the team with Matt Travers, both of CMU’s Robotics Institute. “Our team has a wealth of experience in operating robots in mines, enclosed spaces and the wild, and in coordinating the activity of multiple robots,” Scherer said.
Matt Travers, a systems scientist in the Robotics Institute, said the Carnegie Mellon University (CMU) team will leverage its expertise in modularity. This will allow the scientists to develop robots that can be rapidly built and reconfigured to adapt to widely varied environments. We can’t be sure that a four-wheeled platform will always be the right robot for every job, so we need to be ready to add wheels or substitute tracks or even legs,” Travers said. The kind of robot the team will need depends on the variety of environments. “Small robots might be our only option, while others may demand larger, more robust robots,” he said.
Sebastian Scherer, who is a senior systems scientist, said communications will be a major challenge underground. He added that getting robots to work cooperatively to ensure a space is comprehensively mapped is critical. “Creating robots that can work in subterranean environments will expand the potential application of robots both underground, such as in mines, and inside structures, such as buildings, ships, and aircraft,” Sebastian Scherer said. “The constraints robots encounter in these confined spaces are enormous, so we have our work cut out for us.”
Each team’s robots searched for a set of 20 predetermined objects, earning a point for each find. Collaborative SubTerranean Autonomous Robots (CoSTAR) is developing robots that can autonomously explore caves, pits, tunnels and other subsurface terrain. CoSTAR, which stands for Collaborative SubTerranean Autonomous Robots, brought machines that can roll, walk or fly, depending on what they encounter. Along the way, the bots have to map the environment and find objects like a warm mannequin that simulates a disaster survivor or a lost cellphone with a Wi-Fi signal. This particular course, which aims to simulate an urban environment, also included a carbon dioxide leak and a warm air vent. “The goal is to develop software for our robots that lets them decide how to proceed as they face new surprises,” said CoSTAR’s team lead Ali Agha of JPL. “These robots are highly autonomous and for the most part make decisions without human intervention.”
Joining the team for the Urban Circuit was a four-legged robot called Spot, which was provided by Boston Dynamics. “One of the two courses we had to run had multiple levels, so it was great that the Boston Dynamics robots were fantastic on stairs,” says Joel Burdick, a Caltech professor and JPL research scientist. He is the leader of the Caltech campus section of the CoSTAR team. As the bots explore, they send back video and digital maps to a single human supervisor, who they remained in radio contact with for the first 100 feet (30 meters) or so of the course. They can extend that range by dropping communications nodes, a kind of wireless repeater. Once out of contact, it’s up to each robot to decide whether to proceed or backtrack in order to update the team. Each must also rely on fellow robots to access different levels of the course. For example, a wheeled robot might request a quadrupedal one to climb or descend a flight of stairs.
“These courses are very, very challenging, and most of the difficulty lies in communicating with the robots after they’ve gone out of range,” Agha said. “That’s critical for NASA: We want to send robots into caves on the Moon or Mars, where they have to explore on their own.” Mapping caves on the Moon or Mars could identify good shelters for future astronauts. Moreover, if it exists at all, microbial life has a better chance of survival under the surface of Mars or within the icy seas of planetary moons, like Europa, Enceladus and Titan. NASA wants to search for life in these regions, where robots would be frequently out of contact.
The next circuit in the Subterranean Challenge will be set in an undisclosed natural cave network this August. A final circuit that blends tunnels, urban environments and natural caves will take place in August of 2021. Teams competing in that final event have the opportunity to win up to $2 million in funding from DARPA. CoSTAR, includes JPL; Caltech, which manages JPL for NASA; MIT; KAIST (Korea Advanced Institute of Science and Technology); Sweden’s Lulea University of Technology; and industry partners.
How JPL’s Team CoSTAR Won the DARPA SubT Challenge:
The SubT Challenge is designed to encourage progress in four distinct robotics domains: mobility (how to get around), perception (how to make sense of the world), networking (how to get the data back to the server by the end of the mission), and autonomy (how to make decisions). The competition rules and structure reflect meaningful real-world scenarios in underground environments including tunnels, urban areas, and caves.
To be successful in the SubT Challenge requires a holistic solution that balances coverage of each domain and a recognition of how each is intertwined with the others. For example, the robots need to be small enough to travel through narrow passages, but large enough to carry the sensors and computers necessary to make autonomous decisions and while navigating in perceptually-degraded parts of the course, meaning dark, dusty or smoke-filled. There’s also the challenge of power and energy: The robots need to be quick and energy-efficient to meet the endurance requirements and traverse multiple kilometers per hour in extreme environments. At the same time, autonomous onboard decision making and large-scale mapping is the single biggest power demand. Such challenges are amplified on flying vehicles and require more dramatic trade-offs between flying time, size, and the autonomous capabilities.
Our answer to this call for versatility is to present a team of AI-powered robots, comprising multiple heterogeneous platforms, to handle the various challenges of each course. To enable modularity, all our robots are equipped with the same modular autonomy software, called NeBula (Networked Belief-aware Perceptual Autonomy). NeBula is specifically designed to address stochasticity and uncertainty in various elements of the mission, including sensing, environment, motion, system health, and communication, among others. With a mix of wheeled, legged, tracked, and flying vehicles, our team relies on a decision-making process that translates mission specifications, risk, and time into strategies that adaptively prescribe which robot should be dispatched to which part of the course and when.
The hallmark of Team CoSTAR’s first year leading up to the SubT Tunnel Circuit was a series of fast iterations through potential robot configurations. Every few weeks, we would make a major adjustment to our overall solution architecture based on what we learned in the previous iteration. These changes could potentially be as major as changing our overall concept of operations, e.g., how many robots in what formation should be part of the solution. This required high-levels of adaptivity and agility in our solution development process and team culture.
Three months before the first scored SubT event, the Tunnel Circuit, DARPA revealed that the competition was to be held at a research coal mine in Pittsburgh, Pa. This mine appeared to have less dust, fewer obstacles, and wider passages than our test environments, but it was also more demanding due to its wet and muddy terrain and large, complex layout.
One of the major additions for the Urban Circuit was the introduction of multi-level courses, where the ability to traverse stairways was a prerequisite for accessing large portions of the course. To handle this, we added tracked robots to our fleet. Thanks to the modularity of the NeBula software framework and highly transferable hardware, we were able to go up and down stairs with our tracked robot in four months.
To score even a single point, a chain of events needs to happen flawlessly. Firstly, a robot needs to have covered enough space, traversing mobility-stressing and perceptually-degraded course elements, to reach an area that has an artifact. Multiple camera video streams as well as non-visual sensors are analyzed by the NeBula machine learning framework running on the robot to detect these artifacts. Once detected, an artifact’s location must be estimated to within 5 meters of the true location defined by DARPA with respect to a calibration target at the course entrance. Finally, the robot needs to bring itself back into communication range to report the artifact location within the 60-minute window of mission duration. A critical part of accomplishing the mission in this scenario is a decision-making module that can take into account the remaining mission time, predictive mission risk, as well as chances of losing an asset, re-establishing communication, and retrieving the data. It’s a delicate balance between spending time exploring to find as many artifacts as possible, and making sure that artifact locations can be returned to base before time runs out.
From the beginning, the JPL team forged partnerships with four other institutions offering complementary capabilities to collectively address a daunting list of technical challenges across multiple domains in this competition. In addition to JPL’s experience in deploying robust and resilient autonomous systems in extreme and uncertain environments, the team also included Caltech, with its specialization in mobility, MIT, with its expertise in large-scale mapping, and KAIST (South Korea) and LTU (Sweden), experts in fast drones in underground environments. The more far-flung partnerships were the result of existing research collaborations, a typical pattern in robotics research. We also partnered with a range of companies who supported us with robot platforms and sensors. The shared philosophy of building Collaborative SubTerranean Autonomous Robots led to the birth of Team CoSTAR.
Team CERBERUS wins DARPA Subterranean Challenge, reported in Oct 2021
The final competition of the challenge took place in Louisville, KY last week at the Louisville Mega Cavern, a massive former limestone mine so big it houses a ropes course, mountain bike park, and tram-guided tours.
Finalists in the competition had to navigate environments that incorporated elements from all three previous events, including confined spaces built to simulate underground mines, metropolitan infrastructure, and cave systems. They were faced with visual degradation during operations, including smoke.
To win, robots from each team had to explore these environments and find objects of interest—called artifacts—that had been placed there by DARPA SubT Challenge organizers. Once found, the robots had to report the accurate location of the artifact. Each artifact was worth one point.
Team CERBERUS won by accurately finding and locating 23 artifacts out of the 40 total present. Second place winner Team CSIRO Data61 also scored 23 points but it reported the final artifact more slowly, giving the win to Team CERBERUS. Third place winner was Team MARBLE, with a score of 18.
“Our team coined early on the idea of legged and flying robot combination,” says Dr. Kostas Alexis, Director of the Autonomous Robots Lab at NTNU and Team CERBERUS’ Team Lead. “We have remained focused on this core vision of ours and also bring fully own-developed hardware for both legged and flying systems. This is both our advantage and—in a way—our limitation as we [have spent] a lot of time in its development.”
The DARPA SubT Challenge pushed the limits of autonomy and underground robotics technology. All of the teams that made it to the finals developed and demonstrated technology that has the potential to make a big impact on sectors that rely on underground operations, including mining, wastewater inspections, and search & rescue. The primary robots that Team CERBERUS used in the Final Event were four ANYmal C systems.
“Flyability is honored and proud to have been part of the CERBERUS team, winners of the DARPA Subterranean Challenge,” says Adrien Briod, Co-Founder and CTO at Flyability. “At Flyability we are committed to turning innovations into products that solve customer pains, with the primary goal of using robotics to keep people out of dangerous situations.”
References and resources also include