Data processing: speech signal processing – linguistics – language – Speech signal processing – Application
Reexamination Certificate
1998-11-02
2003-07-15
Abebe, Daniel (Department: 2654)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Application
C704S275000, C704S251000, C704S246000
Reexamination Certificate
active
06594632
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to voice recognition systems, and in particular to improvements in the use of such systems in an environment where hands-free operation is desirable, such as a retail checkout station. The invention relates to advantageous methods of turning a voice recognition system on and off without the need for the system operator to use their hands to push a button or click a mouse, or to issue specific voice commands. The methods and apparatus described in further detail below are well suited to noisy environments where a traditional voice recognition system would have difficulty ascertaining the operator's voice from background noise, and prevent the system from being accidentally activated by a person other than the operator.
BACKGROUND OF THE INVENTION
Retailers have traditionally used a variety of methods to automate the retail customer transaction process. The desire to automate this process is driven by the competing goals of completing the transaction quickly and accurately, while providing the customer with a positive shopping experience and gathering increasing amounts of information about the transaction and the customer.
Retail grocers have been the most progressive type of retailer in regard to automating the checkout process. The near universal implementation of product package barcoding by product manufacturers, and barcode scanning by retailers, has been essential to the implementation of the automation process. One significant problem that remains is the impact that unlabeled products, or otherwise unscannable products, have on the automation of the checkout process. This problem is further compounded by the lack of training and relatively low skill level of cashiers due to rapid employee turnover and other factors. Among the most common reasons for an item being unscannable are that the barcode is damaged or of insufficient quality, the item is not found in the POS system price look-up (PLU) file, or the product package is not barcoded. When a cashier encounters a product that is unscannable, the cashier must typically stop the scanning process and manually take some action to identify the item. If the item has a barcode, but the item does not scan, the operator may try typing the Universal Product Code (UPC) number on the package. If the item consists of unlabeled bulk goods or produce, the operator may consult a paper list to find the proper price look-up (PLU) number. If neither of the above actions is adequate or appropriate, the operator may call for assistance from another employee, enter a fictitious price or just give the item away.
The most common type of non-barcoded product in a grocery store is produce. Although many grocers have implemented the practice of labeling all produce with a PLU number printed on a sticker, individual pieces of produce often lose the sticker by the time the item reaches the cashier. The cashier will either enter a PLU from memory or consult a table of PLU numbers. It is common for cashiers to enter a PLU for a different, but similar, product merely to keep the checkout process moving quickly. This deprives the grocer of accurate sales and inventory replenishment data. Another common scanning problem occurs when the POS system does not find the item in the PLU file. In this case the cashier may ask the customer for the item price, or call for assistance from a supervisor or grocery bagger, resulting in a significant delay in completing the checkout process. If the customer provides the price, the retailer may be deprived of accurate data, and possibly revenue. If the cashier calls for assistance, customer satisfaction suffers while all of the customers in the line endure a delay. Lastly, if the barcode is damaged or otherwise unscannable, the cashier must manually enter the UPC number. If the UPC number is found in the PLU file, the checkout process proceeds with a nominal delay. If the item is not found, the cashier must call for assistance and the customer is delayed further. A cashier experiencing any of the above situations could utilize a voice recognition system to query the POS system for a PLU inquiry, or to voice certain product name or description information to be used in a search query. The problem to be solved, then, is to be able to activate the voice recognition system only when needed, and to do so while continuing to scan other items.
A similar set of problems are present in the food preparation industry. This includes, but is not limited to, the kitchens of a fine-dining restaurant, fast-food restaurant, or grocery store deli counter. Cooks and order takers in such an environment must use their hands to prepare meals or orders, and cannot work efficiently if they must stop their work to push a button or click a mouse to activate a voice recognition system. One major challenge for the kitchen of a fine-dining restaurant is to be able to complete all orders for a given table at the same time, and to notify the appropriate waiters that their orders are ready to be taken to the tables.
In a typical fast-food restaurant, the order takers enter orders into an order-entry computer system. The order-entry computer system then sends the kitchen portion of the order to a computer display in the kitchen. The cooks then remove orders from the kitchen display after they prepare the kitchen orders and ‘bump’ those orders back to the order fillers. The order fillers or order takers gather the non-kitchen portions of the order while the kitchen prepares the kitchen portion. When the kitchen items are ‘bumped’ back to the order fillers, the order fillers take the kitchen portion of the order and combine it with the non-kitchen portion. When the order is completely filled, the order filler delivers the order to the customer and ‘bumps’ the order out of the order-entry system. The ‘bumps’ are typically accomplished by using a bump-bar peripheral device attached to the order-entry system. Utilizing a typical bump-bar requires the order filler or cook to walk over to where the bump-bar is installed and physically push the bump-bar.
The clerks at a grocery store deli counter need to weigh or measure orders before printing a label that identifies and prices those orders. A typical counter will have a plurality of scales that may be used by any one of the clerks. A clerk typically needs to perform a number of tasks to complete an order. These tasks may include filling small containers from large bulk containers in a refrigerated case, cutting a loaf of bread on a bread slicer, and cutting meat products on a meat slicer. After each individual order item is prepared, the clerk must go to one of the scales to weigh the item and enter some identifying information before the scale can generate an appropriate label.
In each of the above situations, the clerks, cooks and order takers could advantageously utilize a voice recognition system to provide the required bumps or data inputs. Utilizing a voice recognition system that is activated in a hands-free manner would allow those workers to provide the necessary data inputs while they were completing their other required tasks.
SUMMARY OF THE INVENTION
The present invention recognizes that there exists a need in a variety of contexts for a hands-free method of starting and stopping a voice recognition system. Such a method should advantageously allow the starting and stopping of such a system without the use of specific spoken commands. Such a method should additionally operate irrespective of the type of ambient noise present where such a system is installed.
Methods and apparatus for operating a voice recognition system in a retail store, according to one aspect of the present invention, preferably include a means of starting and stopping the recognition process without the use of the operator's hands or feet, as well as a means for annunciating the status of the voice recognition process. One method described further below allows for starting and stopping the voice recognition system by blowing on a microphone. Alternate
Abebe Daniel
NCR Corporation
Priest & Goldstein PLLC
LandOfFree
Methods and apparatus for hands-free operation of a voice... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Methods and apparatus for hands-free operation of a voice..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Methods and apparatus for hands-free operation of a voice... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3025750