YOU may think your car says a lot about you, but to a computer called Bob it says just one thing: you like either quarter-pounders or chicken nuggets.
Bob is an automated fast food restaurant management system and its job is to predict what people are likely to order before they reach the counter. The idea is to get the chefs cooking the right food for the incoming customers as soon as possible. It cuts waiting time and wastage, and ensures that the food is hot.
Bob, which was developed by HyperActive Technologies of Pittsburgh, Pennsylvania, has two ways to predict what customers are likely to buy: the type of car they drive and their height. Pickup trucks mean quarter-pounders hit the grill, while big family cars are likely to contain children, so chicken nuggets and fries are the order of the day.
Advertisement
Similarly, people of adult height walking up to the front door are likely to order quarter-pounders, while people under a certain height tend to be youngsters who order child meals.
Bob takes pictures from roof-mounted cameras. They are fed to a PC where image-processing software can identify types of vehicle and sizes of pedestrians. The number of cars also helps predict how many meals will be required, an important factor for the smooth running of a restaurant. The software makes its predictions based on research carried out by HyperActive Technologies. In time, Bob will also use data on past food consumption to help work out future demand.
In a trial at a McDonald’s in Chippewa, Pennsylvania, Bob is making an impression. “Bob is managing supply and demand, right down to the minute,” says Pat Currie, the branch manager. Craig Coulter, co-founder of HyperActive Technologies, says Bob is ensuring burgers are served hot, and waste has been “virtually eliminated”. Waiting times have been cut by an average of a minute per order.
HyperActive Technologies is not alone in trying to predict what people want in restaurants. Advanced Interfaces, a software company based in State College, Pennsylvania, has developed more sophisticated ways of predicting what people will order. Software it is developing will infer what people are likely to order using factors such as gender and age, assessed from a photograph of an approaching customer. “For example, chicken and salads are more popular with females, burgers are more for males,” says Armeen Gould, an applications developer at the firm.
But Coulter counters that Bob is already doing such a good job using only two parameters that there is no need to capture finer levels of detail to accurately predict demand.
Bob is now being tested in seven branches of McDonald’s, a Burger King and a Taco Bell in Pennsylvania, Ohio and Florida.
