Leading Taiwan University Unveils Intelligent Traffic Sensor Software

Drivers who grow drowsy while speeding down the highway need to be alerted to the danger, and Taiwan's elite National Chiao Tung University has developed "intelligent" software, claimed to be the most "human-like" of its kind, to do just that.

Prof. Chang Jyh-yeong of the university's Department of Electrical and Control Engineering Science is in charge of the development project, which is included in the government's four-year "Smart Vision Biological-Inspired Intelligent Vision Technology for Intelligent Transportation Systems," or "ITS" for short.



"This smart vision system reacts just as a human does, making its technology the closest in the field to human biological reaction," the professor claims. To develop the software, Chang and his team used the results of field experiments that studied the blinking behavior of over 10 subjects in a drowsy state. They also employed fuzzy-neuro integration-analysis patterns.

Chang is an expert in fuzzy-neuro systems and image-processing pattern recognition. His system employs a CMOS-based camera connected to the computer that runs the program. The camera sends eye-blink images to the computer, which processes them and analyzes the blink patterns; when the blinking frequency slows down to a drowsy state and eyelid coverage of the eye increases, the program commands the computer to issue a sharp audio warning—a beep—to alert the driver.

The ITS project is being carried out by National Central University and the National Taipei University of Technology, as well as Chiao Tung University, which is in charge of program coordination and personnel supply. For each of the past three years, the three universities together have been granted NT$40 million to NT$50 million (US$1.3-1.6 million) in subsidies for the program from the Ministry of Education (MOE). The project will come to an end in March next year.

Traffic Warning Too

The ITS has other functions in addition to warning sleepy drivers. When linked to a global positioning system (GPS) receiver and a web camera, it can, with certain modifications, allow drivers to check out real-time traffic conditions on nearby roads via a wireless portable computer. A demonstration area for this system has been set up to cover the Chiao Tung University campus, the nearby Hsinchu Science-based Industrial Park, and part of Chupei city.

Installed with a home-security system, the intelligent-sensor system can send out alerts as soon as it locates intruders and analyzes their digital images. "CCTV sensor systems have to be watched by duty personnel at all times," Chang comments, "but our intelligent system does not."

Taiwan's well-developed electronics industry gives the island an advantage in developing intelligent transportation systems, which, the professor says, are still in the fledgling stage in many industrially advanced countries. "With the lackluster business of the computer industry in the past few years," Chang notes, "information-technology device manufacturers are trying to capture a bigger slice of the market by adding new applications to their business scope. Transportation is a promising application.

Local enterprises in Taiwan have approached the development team about the transfer of technology, which, Chang claims, "is the first of its kind in the world to be ready for commercialization. Eventually, we hope to develop it so that it can use a nail-size microprocessor instead of a desktop or notebook computer."