Next spring Sharp, the electronics company, will launch a microwave oven with the brains to work out the time and power level needed to reheat any food, regardless of weight or starting temperature. All the user has to do is switch it on.
A neural network is a computer programme that learns by experience, like a real cook. The Sharp neural network perfected its art in the laboratory by being exposed to the reheating of countless samples of food. The microwave oven has a humidity sensor that detects the water vapour released as the food is heated. Based on its 'previous' experience in the lab, the neural network recognises food characteristics such as temperature and quantity from the humidity levels.
The time and power needed to reheat a particular food is not just a function of starting temperature and weight, but also varies with the type of food. To work out how to reheat a particular type of food, the neural network assesses the way the humidity level changes over time. Rice, for example, is ready to eat at a lower temperature than its accompanying curry sauce. The oven works as a conventional microwave for raw food.
It was developed at Sharp's laboratories in Oxford in collaboration with Oxford University's engineering department. The project began in response to market research showing that most people in the UK and Germany use microwaves for reheating rather than cooking raw food, and wanted a single- button operation.
The oven will be made at the company's plant in Wrexham for the European market. Last week Sharp announced a pounds 7.8m investment in microwave oven production there.
The neural network oven is the first Sharp product to be developed outside Japan. The Oxford lab, which opened in 1990, does original research in a range of other technologies, including liquid crystals and optoelectronics, but the rest of its work to date has been carried out under the direction of the company's Japanese lab.Reuse content