Robots

The assigned article mentioned how incredibly difficult it was for a robot to autonomously map its surroundings and understand its position within them. A human being is able to simply open his/her eyes, look around at the surroundings and can recognize landmarks and other important indicators of his/her whereabouts. The extraordinary difficulty for a robot to complete a very basic human task interested me and was the basis for me when looking through articles for this week’s discussion.

The first article is an experimental summary from the late 1990’s titled “A Robot Map Creation Algorithm.” This article provided both the logic and experimentation behind an early attempt at constructing an algorithm for a robot to both construct a map of its surroundings and localize itself within them. The theory for this process was simple enough and made sense, the robot would drive around to 28 different locations on a map and would take sonar readings at each location. These sonar readings would essentially be pictures for the robot to study but would have a unique imprint for each one. Then, once the robot completed its journey to the 28 stops, it would compile the sonar readings it took into a sonar map. Finally, to test the robot’s ability to understand its surrounding, the robot would be placed at any of the locations, would take a sonar reading, compare it to the stored sonar map and would output which location most directly related to its current location. The author mentioned that some initial problems included sonar readings of the actual robot as some “stealth” readings, where the sonar would essentially bounce off angled surfaces. The robot was programmed to discard these outliers so as not to interfere with the actual sonar readings. In addition, the author mentioned that there were some errors with the actual sensor and that the reported distances varied somewhat from the actual distances. However, I would expect that since the article was posted, over ten years ago, that the technology in sensors has increased greatly.

The second article dealt with the application of these autonomous vehicles and the possibility of an autonomous highway. The article started out by describing how the growing traffic issues have been an ongoing problem and have been calculated to cost the US economy alone over $50 billion. The past solutions to this problem of expanding highways and making more lanes would be nearly impossible given the densities of the areas where these highways are in. However, making traffic flow more efficiently would be able to solve the problem. The combination of autonomous vehicles, which can be programmed to drive at a very consistent pace with a very close proximity to other cars, and a complex matrix of computers would allow for more capacity on the current highway lanes. An increase from 2,000 to 6,000 cars per hours passing could be expected from this improvement. This increase is only possible because of the creation of robots. Without the precision and ability for several cars to drive within a few feet of each other at high speeds, the current lane capacity would remain fixed. The two second rule for judging an appropriate distance to allow for the car ahead of you could now be replaced with a standard two foot gap, regardless of what speed the cars are traveling. This article really provided a great application of the robotic technologies and showed how the world as we know it may likely change in the next few years!