Technology Bulletin 1107

Neural-net software may fly crippled aircraft

NASA has two experimental programs under way that use neural-net software to help control damaged planes, as well as help them land. The generic control software could eventually be used in many commercial industries, according to its developers. Neural-net software learns by observing pairs of related patterns in the real world, and then performs different tasks in response to these patterns. It consists of many interconnected processors (or nodes) using computational principles derived from neurons in the brain. Each node assigns a value of the input from each of its counterparts. As these values change, the network can adjust the way it responds. The new software is under joint development by scientists at NASA's Ames Research Center and McDonnell Douglas. It may allow a damaged aircraft's computer to "relearn" to fly and then land a plane safely after a major equipment failure or explosion. E-mail [email protected].

USCAR successfully crashes composite structure

American automakers reached a milestone in developing advanced, environmental technology by meeting government crash-safety standards with a lightweight, composite front-end structure designed for mass production. The test, performed by the U.S. Council for Automotive Research (USCAR), involved a Ford Escort fitted with upper and lower composite front-end rails that acted as the primary load-bearing structure for the engine and bumper. USCAR manages collaborative research of Chrysler, Ford, and General Motors. The accomplishment marks the first time the government safety criteria for a 35-mph frontal barrier crash test were met with a production-feasible composite design. The composite parts, made of polymers reinforced with glass fibers, weigh 25% less than the steel parts they replaced on the test vehicle. "Working together, we were able to develop state-of-the-art technology suggesting composites can be designed to function in crashes as well as traditional steel," says Mark Botkin, USCAR's Automotive Composites Consortium's project leader and GM principal research engineer. FAX (313) 248-4303.

'Smart airbags' would help reduce accidental deaths

The Optoelectronics Group of EG&G Inc., Wellesley, MA, has under development a new technology it feels could dramatically reduce the number of accidental deaths and injuries caused by automotive airbags. Current airbags that deploy at speeds approaching 200 mph, while protecting an adult, can be crushing to a child. EG&G's German-based Heimann division has already developed sensors that employ an infrared array to detect whether a person is sitting in a car's passenger seat or not. By reading a person's body heat, these sensors can detect the person's head position. They can, for example, differentiate between a person and a large bag of groceries. Simultaneously, a second ultrasonic sensor transducer can detect the distance a person is sitting from the bag's cover. The EG&G sensors would automatically deactivate airbags if a small child was sitting in the passenger seat, or the driver was too close to the airbag. FAX (617) 431-4255.

Digital speech analysis could identify drunk drivers

Slurred speech is often a telltale sign that someone's been drinking. Now, a Georgia Tech researcher is at work with colleagues from Indiana University to digitally quantify this trait, which could lead to a simple, non-invasive way to test a person's sobriety. "It's basically an effect of fine motor control," explains Kathleen Cummings, an instructor at Georgia Tech's School of Electrical and Computer Engineering. Preliminary results show that intoxicated speech is marked by jumpy changes in pitch and energy production, as well as unsteady opening and closing of the vocal cords. Cummings says translating her research into a practical public-safety device could be relatively easy. With it, law-enforcement officials could record someone's speech at an accident or traffic stop, then analyze it later by computer against a sample taken at a different time. E-mail [email protected].

Ozone laundry system destroys germs, cuts energy costs

The Electric Power Research Institute (EPRI) and Tri-O-Clean Systems, Fort Pierce, FL, have introduced to the healthcare industry an innovative laundry system based on the use of ozone. The system incorporates the highly reactive form of oxygen in the wash water to clean more quickly and thoroughly with less energy, water, and washing chemicals. "Ozone is 15 times more powerful a disinfectant in water as chlorine and many times faster in destroying viruses and bacteria," says Myron Jones, EPRI's project manager for the program. In the system, an onsite ozone generator produces ozone via an electrical discharge. The ozone gas is then mixed with water flowing to the washing machine, where it transforms soil from insoluble to soluble, enabling it to be loosened and removed from fabric. Lint and other particles are removed with the wash water through a series of filters. The recycled water is then re-ozonated to be used in the next wash cycle. E-mail [email protected].