Image Credit: Anurake Singto-on/Shutterstock.com
Rescue teams risk their lives every day to save people during fires, floods, earthquakes, chemical accidents, and in bomb disposal work; robots and drones have a proven track record in providing assistance in such search and rescue operations where sending in humans is too dangerous.
Initially developed for the military, these machines navigate situations that are hazardous for humans and provide critical lifesaving support, whether that be finding victims, diffusing bombs, or bringing supplies to those who are trapped.
While they work well, scientists and engineers are still looking to improve today’s robots, aiming to make them more maneuverable and more intelligent using artificial intelligence (AI) and machine learning.
Robots were first employed in search and rescue operations in an urban environment during the September 11 attacks in America in 2001. It can take five to eight hours to inspect a collapsed building before searching for victims – the use of unmanned aerial vehicles (UAVs) and drones were able to shorten the process by assessing the structure of the buildings, searching for paths through the rubble and locating victims.
Although they provided the necessary quick response and assessed hazardous conditions which could have put human life at risk, the robots could not penetrate the complex depths of the building because of the limited mobility and intricacies of guiding with a joystick.
Since then, robots have been used as remote-controlled tools to help locate buried accident victims, defuse bombs and dismantle decommissioned nuclear power plants. Great advances have been made in the last 18 years, robots are now equipped with autopilot systems and vision systems capable of identifying dangerous situations and recognizing people. They are also loaded with sensors to pick up geometrical features and humidity levels.
The Big Challenge
The challenge is to develop unmanned rescue to robots capable of making independent decisions that can work unsupervised in a small, chaotic area.
The maneuverability of robots in confined spaces needs to be improved, as does their self-awareness, which they need to recognize the intent of the mission as conditions change. They need to be able to adapt their locomotion style automatically depending on whether they need to walk or crawl, swim, roll, fly, or climb. In the future, robots may also need to be able to alter their shape, geometry, and movements based on the terrain or task at hand.
Systems need to gain greater independence; they need to be able to alter their operations as they gather new information while also following and cooperating with human operators. They require improved AI to allow robots to learn to solve problems without data or human experience; enhanced AI will create robots that can move with minimal human assistance and self-adapt to situations.
The military, in particular, are interested in developing highly maneuverable flying UAVs with robotic arms that can navigate highly confined spaces and interact directly with its surroundings, taking samples, moving debris or providing medical assistance for example.
Equipping robots with artificial intelligence is central to communication and collaboration between humans and machines. In specialist groups at the German Research Center for Artificial Intelligence, experts on language processing and industrial informatics are teaching robots to understand rescue operation procedures, comprehend commands, and provide feedback to rescuers. The aim is for natural and efficient communication with robots so they complement human emergency workers.
In America, Sandia Labs have created Gemini-Scout, a rolling rescue robot designed to make mining rescue missions easier by replacing humans in the early stages of rescue planning and response. Featuring heavy-duty wheels, hi-tech surveillance equipment, and payload storage, it can evaluate hazardous conditions, enter potentially risky environments, provide feedback, and deliver supplies to miners who need rescuing.
Engineers from JPL Robotics have developed RoboSimian, a robot equipped with four general-purpose limbs that can move and rotate. This means the robot is mobile and can manipulate itself to respond to the crisis at hand. It uses careful and stable operations to carry out challenging tasks, with the emphasis on steadiness over dynamics and deliberation overreaction for faster and more robust operations. This means that the robot can achieve passively stable stances, establish multi-point anchored connections to supports such as ladders, railings, stair treads and to brace itself during forceful manipulation operations.
Unmanned ground, submarine, and aerial vehicles and robots can save lives by responding to a disaster, both natural and manmade, faster and containing emergency situations more quickly than traditional techniques and tools.
Sources and Further Reading
- Intelligent robots save lives - Deutschland
- 3 Robots That Are Designed to Save Human Lives - Tech co
- DRC - DARPA Robotics Challenge, RoboSimian (Track A) - Robotics JPL
- Robots to the Rescue: Saving Lives with Unmanned Vehicles - the Conversation