ROBOTICS research progressing in Qatar could contribute significantly in making cars that drive themselves a reality for the masses in the not too distant future, ushering in unprecedented traffic safety.
Carnegie Mellon University in Qatar (CMUQ) is anchoring an initiative in this direction, with graduate student in robotics, Justin Carlson’s doctoral work.
“The specific topic I am working on is integration of global positioning system (GPS) with simultaneous localisation and mapping (SLAM),” explained the robotics researcher, who did his thesis proposal this February.
If GPS is a widely used satellite based navigation aid, SLAM is a technique used by robots and autonomous vehicles to build a map within an unknown environment while simultaneously tracking their own position.
Though SLAM has not yet been fully perfected, it is being employed in unmanned aerial vehicles, autonomous underwater vehicles, planetary rovers and newly emerging domestic robots.
“I have been experimenting with this technology out in the streets,” said Carlson whose sport utility vehicle is fitted with a laser scanner, a GPS unit, and a computer to record relevant data and create maps.
Recently, he drove around an area close to the Immigration roundabout where there are streets in a grid pattern, generating a map that could be used as an input for autonomous vehicles.
Deploying either GPS or SLAM alone would not do the job as both have their own merits and demerits, but they do complement each other.
Carlson said that a high end GPS can do a pretty good job, within half a metre, as long as there is unhindered access to the sky, whereas SLAM does best when there is a bunch of clutter around, because it works off landmarks.
“GPS tells you where you are, but not accurately enough. It also has problems with availability, particularly when you are working, for example, in the area down by the City Center, where the tall, reflective buildings really mess up the signal,” he observed.
The researcher has also found that conversely SLAM does terribly on some of the major roads in Qatar, especially those with wide open areas around.
“So hopefully, we get some reliability in that GPS would work where SLAM does not and vice versa,” maintained Carlson who aims to do larger maps, probably the size of Doha city itself.
The youngster, engaged in robotics research at CMUQ since 2004, has the goal of enabling a good number of autonomous cars in a place like Doha, which is facing a growing problem of traffic congestion.
“Cars that drive themselves or those that communicate with each other is no longer science fiction,” he remarked while citing the example of Boss, the CMU Pittsburgh Tartan Racing Team’s autonomous vehicle that won the $2mn 2007 US ‘Defence Advanced Research Projects Agency Urban Challenge’.
Features such as adaptive cruise control and lane assist, found in some high-end luxury cars, are results of previous research, Carlson pointed out, before predicting that these would become standard in all cars in the foreseeable future, just as automatic transmission was confined to luxury cars in the beginning.
The researcher is of the view that Qatar is a reasonable place to introduce the first legal, autonomous car on the road, as he believes the authorities here maybe more willing to try something new, in contrast with the US where people are very nervous about new things.
Carlson also spoke about the challenges he is facing in his experiments with the ‘mapping’ project. “The laser scanner, mounted on the roof of my car, gets ‘confused’ when there are a bunch of cars moving alongside,” he said.
For example, if there are taller cars around and all vehicles are moving forward at the same pace when a signal turns green, the laser scanner would think Carlson’s car is not moving.
“Or even worse, if I don’t move when everyone else moves, the scanner would say I have gone backwards. Congested areas are more difficult to deal with,” he maintained.
However, Carlson intends to tackle these issues with some interesting techniques discovered by other researchers to deal with moving objects and SLAM.
Yet another problem he is facing in Doha is that landmarks keep changing as buildings come up and are taken down.
“So you have to kind of put in some intelligence on the lines that landmarks don’t always exist forever. The robot should be able to update maps as the vehicle pass through each area,” he added.
Follow us on our social media channels: