Abstract
The intention of the final contest is to demonstrate the capability to build robots to complete a Search and Rescue mission. In order to complete the mission, the robots must collaborate in order to find victims and supplies in the field and return them to the hospital area.
Why Search and Rescue?
Search and Rescue is an important, real-world task with potential to help people who become lost or stranded in a disaster, or anyone who is in a situation where they need help and no one else is around. By simulating Search and Rescue in the lab, we learn the difficulties involved in developing robots to perform similar tasks. Because Search and Rescue is a task that one day may be done effectively and more efficiently by robots it is useful as a robotics project.
Why Robots?
Robots, at least theoretically, can accomplish many tasks just as well as or superior to humans. Although robotics technology has not advanced to the point of replacing very many human activities, robots have already taken over some dangerous tasks from humans, including some factory work, exploring volcanos, and even exploring the nuclear accident at Three Mile Island. Perhaps in the future robots can take over not only dangerous tasks, but tedious and time-consuming tasks as well. In the case of Search and Rescue, the task can be dangerous because it is so slow and tedious. Searching over a wide area for a relatively small victim can take hours, days, or even weeks, and in the tedium of the search the victims might not be found fast enough, or even be passed over.
In addition, there tend to be only a limited number of people who have the training and experience to search with high effectiveness, and even these people would need rest, shelter, and supplies, reducing the time spent searching. The task of searching can be dangerous in itself, with vehicle accidents, helicopter accidents, accidents due to fatigue, etc. Robots could search longer hours, faster, and with more effectiveness, finding the victims faster, and accidents involving damage to robots are preferable to accidents involving damage to humans.
Why is robotics hard?
Why are robots not doing these tasks already? The tasks of localization (finding where the robot and victims are), collaboration (coordinating with other robots and the home base), and vision (using cameras as a sensor) are already nontrivial in a lab setting, so outside in the real world they become impossibly difficult. The real-world is large, detailed, dynamic, noisy, and generally unfavorable to robots. There are many unsolved problems in robotics relating to these areas which are necessary but generally unrelated to the specific task at hand. Outside of the controlled environment of the lab, many problems including sensor unreliability, environmental dynamism and others are encountered which affect the performance of the robot at the task.
Why in the lab?
Solving the complete task of building an effective Search and Rescue robot that can outperform humans is perhaps not feasible using today's technology, and certainly impossible in the span of a semester. Therefore, by selectively choosing pieces of the problem, we can begin to solve those problems which in turn solve part of the complete problem. The techniques used, though insufficient to solve the complete, real-world problem, can be used as a foundation in continuing work that one day may infact build such a robot that can perform the complete, real-world Search and Rescue task.
The intention of the final contest is to demonstrate the capability to build robots to complete a Search and Rescue mission. In order to complete the mission, the robots must collaborate in order to find victims and supplies in the field and return them to the hospital area.
Why Search and Rescue?
Search and Rescue is an important, real-world task with potential to help people who become lost or stranded in a disaster, or anyone who is in a situation where they need help and no one else is around. By simulating Search and Rescue in the lab, we learn the difficulties involved in developing robots to perform similar tasks. Because Search and Rescue is a task that one day may be done effectively and more efficiently by robots it is useful as a robotics project.
Why Robots?
Robots, at least theoretically, can accomplish many tasks just as well as or superior to humans. Although robotics technology has not advanced to the point of replacing very many human activities, robots have already taken over some dangerous tasks from humans, including some factory work, exploring volcanos, and even exploring the nuclear accident at Three Mile Island. Perhaps in the future robots can take over not only dangerous tasks, but tedious and time-consuming tasks as well. In the case of Search and Rescue, the task can be dangerous because it is so slow and tedious. Searching over a wide area for a relatively small victim can take hours, days, or even weeks, and in the tedium of the search the victims might not be found fast enough, or even be passed over.
In addition, there tend to be only a limited number of people who have the training and experience to search with high effectiveness, and even these people would need rest, shelter, and supplies, reducing the time spent searching. The task of searching can be dangerous in itself, with vehicle accidents, helicopter accidents, accidents due to fatigue, etc. Robots could search longer hours, faster, and with more effectiveness, finding the victims faster, and accidents involving damage to robots are preferable to accidents involving damage to humans.
Why is robotics hard?
Why are robots not doing these tasks already? The tasks of localization (finding where the robot and victims are), collaboration (coordinating with other robots and the home base), and vision (using cameras as a sensor) are already nontrivial in a lab setting, so outside in the real world they become impossibly difficult. The real-world is large, detailed, dynamic, noisy, and generally unfavorable to robots. There are many unsolved problems in robotics relating to these areas which are necessary but generally unrelated to the specific task at hand. Outside of the controlled environment of the lab, many problems including sensor unreliability, environmental dynamism and others are encountered which affect the performance of the robot at the task.
Why in the lab?
Solving the complete task of building an effective Search and Rescue robot that can outperform humans is perhaps not feasible using today's technology, and certainly impossible in the span of a semester. Therefore, by selectively choosing pieces of the problem, we can begin to solve those problems which in turn solve part of the complete problem. The techniques used, though insufficient to solve the complete, real-world problem, can be used as a foundation in continuing work that one day may infact build such a robot that can perform the complete, real-world Search and Rescue task.
