Making Sense of Traffic

By Gayathri Vaidyanathan

Imagine that your car gets stuck in traffic in India. As you wait, small two-wheelers go around you and move to the front of the line. Auto rickshaws navigate expertly through gaps you thought were non-existent. And, finally, a truck rolls to a stop next to you in the blistering midday heat. What would you give to avoid this?

In developed nations, where lane discipline is followed and traffic density is relatively low, it is easy to monitor traffic conditions using sensors. But in a nation like India, with overstretched infrastructure and a ballooning, heterogeneous vehicular population, the scenario is very different. Researchers at Microsoft Research India are working on a solution to this, though.

Venkat Padmanabhan, Ram Ramjee, and Prashanth Mohan of the Mobility, Networks and Systems (MNS) group, are working on developing software for smartphones—cellphones that are capable of downloading and running applications—that automatically detects current road and traffic conditions. These phones would use hardware already present in phones, such as cameras, GSM and Bluetooth radios, microphones, and accelerometers to interact with the physical world. “We expect that in the future a large number of phones, even at the lower end of the market, will feature a number of these sensing capabilities,” says Padmanabhan.

Venkat Padmanabhan

This project is part of a larger exploration by the MNS group of the research possibilities offered by the abundance of sensors of various types in today’s mobile devices.

“The vision is to leverage smartphones that people generally carry to monitor traffic conditions in a city,” Padmanabhan says. “So if you keep the phone in your pocket and drive, that phone will monitor the traffic conditions and report that information to a server.” The aggregated information would provide a rich picture of road conditions.

The system excels in two aspects, he says. The use of the ubiquitous cellphone means that no special infrastructure needs to be installed for monitoring traffic.

“This system,” Padmanabhan says, “allows traffic monitoring purely through software and involves no additional costs.”

In India, where cellphone use has exploded, this system makes a lot of sense. In 2007 alone, more than one billion cellphones were sold worldwide, and a large number of these were smartphones with programming capability. A significant proportion of these sales were in the developing world, including India and China.

The second point that sets the system apart from its competitors is its ability to navigate through unpredictable road conditions in developing nations. Current monitoring systems use sensors and cameras to detect an even flow of similar types of vehicles. For example, in North America, Microsoft introduced ClearFlow, which uses sensors to predict road conditions based on time of day, weather conditions, and other variables, thereby helping drivers avoid traffic jams.

But in India, where highways are few, side roads may be unmapped, and traffic is notoriously unpredictable, such a system would be impractical.

Ram Ramjee

“In cities of developing nations, such as Bangalore, road conditions vary,” Padmanabhan notes. “There are potholes and bumps in the middle of the road, there are many different types of vehicles on the road, and the flow of traffic is not smooth. There are trucks and, at the same time, two and three-wheelers.The rates of progress of these vehicles differ; the smaller vehicles move through faster, the big ones get stuck, and there is also a lot of braking and honking.”

In such a situation, the Nericell system evolved by Microsoft Research India will use the cellphone in your pocket to transmit data using its various built-in features. The microphone would pick up honking, and the accelerometer would detect braking patterns. The GSM radio would triangulate your position using GSM towers so that your approximate position could be determined. And potentially, the cellphone’s camera could capture real-time images. All these would enable the annotation of maps in real time and assist users make driving decisions.

“For instance, Nericell could help find the route,” Padmanabhan says, “that gives users the least stressful drive.”

Collection of such real-time data raises questions, though. As Ramjee says, “The key challenge is to minimize energy usage of the phone—and also respect privacy of the users.”

Prashanth Mohan

Mohan says that, to avoid draining power, the system works using to GSM radio locate the user instead of GPS and also works with ”triggered” sensing, ensuring optimal usage of the phone’s power. Only the low-power accelerometer and the GSM radio needed for phone operation are on all the time.

“When there is a lot of braking happening, as in a busy intersection,” Mohan adds, “the more energy-draining components such as the microphone would be triggered to be turned on.”

The researchers have developed a prototype based on smartphones running the Windows Mobile software. They are planning to test the system later this year. If the prototype works and moves into product development, you might soon be able to avoid those messy streets and intersections.