 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
Engineering Safer Planes
In November of 2001, a plane leaving New York's JFK International Airport lost its tail shortly after takeoff and plunged into a neighborhood below. On a partly cloudy February morning in 1996, an airplane carrying 82 passengers landed "wheels up," or without its landing gear deployed, at Houston Intercontinental Airport. In both of these incidents, the National Transportation Safety Board laid most of the blame on human error; however, both cases also involved potential problems with the plane itself, which, when combined with the human mistakes, led to the accidents. In one case, the flight control system, while functioning properly, may not have offered the pilots the detailed information they needed to avoid trouble. Thanks in part to the ongoing work of an LSU research team, such problems and incidents may soon be a thing of the past. The efforts of LSU Professor Jorge Aravena and a group of electrical and computer engineers could eventually lead to airplanes that will be able to detect problems sooner and provide pilots more specific information on how to respond, or, if need be, take appropriate actions on their own. The project is part of the Aviation Safety and Security Program, a joint research venture through NASA and the Federal Aviation Administration, and it includes researchers from the University of New Orleans (UNO) and the University of Louisiana at Lafayette (ULL). The team recently wrapped up a round of work on the development of flight control systems that was supported by a three-year, $1.6 million NASA and Louisiana Board of Regents grant. Last fall, the researchers began to finalize the results of the initial research and found out that NASA had approved a two-year extension of funding for their work. With this approval, more than $1 million in additional funding will be provided to the researchers. Aravena explained that each university group has certain areas of responsibility. At LSU, he and Kemin Zhou, chair of the Department of Electrical and Computer Engineering, lead a group responsible for the development of "fault tolerance strategies" that are capable of preserving the integrity of the plane even when problems have occurred. The ULL and UNO groups are developing different approaches to fault detection in aircrafts. "Control systems are normally calibrated for certain operating situations," explained Aravena. "If things vary too much, the system might not work, or may give 'bad advice.' Pilots may err as well." Simply put, Aravena said that the university teams are looking for "new ways of detecting faults and dealing with them." Currently, most commercial aircraft are fully automated during normal flight conditions. Pilots monitor the automated functions of the flight-control computer from the cockpit. If there is a malfunction, the pilots call on their knowledge and training in order to take action to correct the malfunction. The goal of the LSU team and other researchers is to develop a better aircraft self-diagnostic in which the plane's computer system detects a problem earlier, recommends the best course of action or, if necessary, takes control of the flight. This "automatic safe recovery" is implemented through novel techniques first proposed by Zhou. To help with their mission, the LSU researchers are using software developed by NASA, including a mathematical model of an entire Boeing 747. Using the Boeing model, Aravena and Zhou are able to simulate how the plane may react in specific situations. They and the other team members brainstorm different situations in which the plane may be jeopardized and can test the automatic safe recovery strategies. "Theory can only take you so far," Aravena said. "And the model of the B747 is the closest thing we have to a laboratory." Aravena described himself, Zhou, and the various graduate and undergraduate students who are working on the project as "control people," not aerospace experts. Therefore, he said, they had to spend a good deal of time learning about planes and their special instruments and requirements. He said that one possible result of their work is that other "control people," such as instrument manufacturers, could glean valuable information and improve their technology or develop entirely new instruments.
In addition, he said that the tools or systems that they develop could be applied to other processes. For instance, industry could adapt the fault detection/control technology for use in plant safety. Aravena said that the teams submitted progress reports to NASA in March 2004, citing numerous positive results. This report was a major factor in nailing down the two-year extension, he said. The teams will continue to build on their discoveries and, eventually, the researchers from all three universities will integrate their results into a computerized "test bed" simulation--a test as close to real-life as possible. Then, the results will be provided to NASA, which will thoroughly test them in its own sophisticated laboratories. Assuming all goes well, the third stage will be to install the new equipment on an experimental plane to see if it works. Finally, Aravena said, the "toughest" step will come when they must prove to the Federal Aviation Administration that the devices they have created cannot possibly cause an accident. "It's a long road, but, in 10 years or so, it could lead to the next generation of revolutionary planes," said Aravena. Contact Rob Anderson | LSU
University Relations Related Links LSU Department of Electrical and Computer Engineering |
