01-Oct-2021 | Market Research Store
The upgraded path planning algorithm for determining ideal paths not only finds routes three times faster but also takes less processing time. The technology allows robots to assess the terrain's difficulty before determining a successful way forward, which may entail bracing against a wall with hands while stepping ahead.
The advanced algorithm developers employed machine learning to teach the robot different approaches to assist its locomotor apparatus in order to maintain balance and proceed. The robot can utilize the pre-learned approaches to evaluate the difficulty based on the obstruction throughout a path and discover an appropriate route to reach the destination much faster. Geometric representation of the entire map will be required to execute this route-determining procedure. Practically, a flying drone would be required that flies ahead and guides the robot throughout the journey.
The team's solution surpassed traditional ways in terms of both success and overall time to plan in a trial using a humanoid robot traveling in a corridor of rubble, which is critical when an immediate response is required during critical scenarios. After conducting more than 50 trials, the researchers found that the revised algorithm-based robots were 84 percent efficient and took only two minutes to discover and process routes. On the contrary, the robots loaded with the preceding algorithms required over three minutes to plan the basic path.
The scientists also demonstrated the capacity of their technology to run on a realistic mobile manipulator, which is a wheeled robot equipped with two arms and a torso. The robot's base was put on a steep slope, so it had to brace itself on an uneven surface with its arms as it traveled. The robot used the team's approach to identify the best possible route in less than a tenth of a second, compared to more than 3.5 seconds with the standard path planner.
The team is planning to integrate dynamically stable motion, akin to human and animal natural movement, which would set the robot free from constantly maintaining the balance and potentially boost its velocity of movement.
Read Also: