Warehouse Management Voice Technology FAQs

Accord Voice Technology – Frequently Asked Questions (FAQ)

Click on Question to view Answer

Voice technology, as used in the warehouse, involves the use of a wearable computer with a headset and microphone so that warehouse operatives receive instructions by voice and verbally confirm their actions back to the system. The wearable computer communicates with the Warehouse Management System (WMS) via a radio frequency (RF) local area network (LAN).
The most common application is Voice Picking of customer orders, where improved accuracy and productivity offer a fast payback, but the Accord warehouse management software (WMS) also enables you to use it for Goods Received, Pallet Put-away and Letdown, and Stock Checking.
Accord uses Vocollect's Talkman T2, T2X or T5 terminals. Talkman® is the industry's most advanced voice-directed, wireless, wearable computer hardware. The belt-worn terminal utilizes advanced speech recognition and speech synthesis technologies, providing users with real-time radio communication to the Accord Warehouse Management software (WMS). You will need an 802.11b compliant radio frequency (RF) network and a PC on the network to act as a communications server.
Provided your network is 802.11b compliant, with 11 Mbps or 2 Mbps access points, it is likely you can use the existing network. It will be necessary to conduct an RF survey to check RF coverage in the warehouse because the radio reception of a belt worn unit can differ from that of hand held or truck mounted terminals.
Improvements in order picking accuracy are dramatic, and accuracy of 99.9% (one error per thousand picks), and often much better, is usually achieved. The improvements that you obtain will depend on your current method of order picking, but if you are moving from a paper-based system to voice directed picking, picking errors are usually reduced by between 80% and 90%.
Warehouse picking productivity usually improves by 10% to 20% because the hands free and eyes free operation speeds up picking, and trips back to the assignment desk are eliminated. Administrative productivity is improved because the work of printing and distributing picking lists or labels is eliminated, as is the task of keying picking confirmations, picking amendments and catchweights into the WMS.
Eliminating paper picking labels brings a significant cost saving in the cost of the paper alone.

The real time radio communication enables real time stock updating. This in turn allows the triggering of letdowns to replenish picking faces, streamlining warehouse management by optimising the use of fork lift trucks and preventing re-picks or waiting time due to empty picking faces. Cycle counting can built in to the replenishment (letdown) task, improving the efficiency of the stock checking process. The improved accuracy of stock recording leads to improved service level and less time spent investigating stock discrepancies

Safety is improved as the hands free and eyes free operation leads to fewer accidents. Eliminating paper also leads to less waste paper or label backing sheets, resulting in a cleaner, tidier and safer warehouse.

The training time for new pickers is reduced by the use of voice, as a voice directed task is easier to learn than interpreting a paper task. Training time can often be reduced by as much as half.

If you operate more than one picking shift then the costs are shared across picking shifts and payback can be achieved in as little as 6 months. If you operate a single picking shift, then a payback of one year is more realistic.
Yes, we will ask you a few questions about your current operation and help you calculate an ROI for the system. Alternatively, you can find a payback calculator online.
Yes, voice technology is ideal for use in the freezer, and the hands free operation offers even greater productivity improvements as gloves hamper the use of paper or radio data terminals. The Talkman T2 terminal operates in temperatures down to minus 29°C.
Yes, the Talkman terminal uses Vocollect's next-generation speech recognizer, BlueStreakâ„¢, which was specifically designed for noisy warehouse and industrial environments. BlueStreak easily accepts long, rapidly-spoken digits, as well as single words, with precision and accuracy, while filtering out and rejecting intermittent and background noise.
Talkman uses speaker-dependent voice technology. Speaker-dependent systems require each user to train the system for his or her individual speech pattern, dialect, or language. Voice training takes only ten to fifteen minutes per Talkman user. Training allows the system to be much more accurate and robust in an industrial environment where there is a wide range of accents, dialects, and languages, and much extraneous noise.

Speaker independent technology is designed to match the speaker's voice to previously created, generic voice patterns, and is less suitable for the warehouse environment.

It is often sufficient for a user to perform the voice training just once, and it takes only ten to fifteen minutes. However, during voice training users do not always speak as they normally would when out in the warehouse. If Talkman consistently fails to understand a particular word the user can retrain their voice on the fly in the middle of a task by simply pressing a button on the unit and speaking that word once more.
Yes, each user trains the system for his/her own voice allowing Vocollect's speech recognition technology to accurately recognize any user's speech regardless of accent or dialect. A user can even train the system to recognise responses in a different language.
You need enough units to cover the number of workers on the biggest shift, plus a small percentage of spares. However, users usually have their own individual belts and headsets.
The time a Talkman unit can operate on a single battery charge depends on a number of factors. The amount of power consumed by the radio card varies significantly between different radio cards, the cold temperatures in a freezer reduce battery life, and the amount of radio communication needed by the application also affect power consumption. Typical times are between 4 and 8 hours. In ambient temperatures, a battery charge may last a whole shift, depending on the other factors. Operation with a faster RF network in a freezer would almost certainly require a battery change during a shift. Standard and heavy duty batteries are available.
Yes, the Talkman terminals are designed to be rugged and have been drop tested with multiple drops of 5 feet onto concrete.
The Talkman weighs less than a pound, including a standard battery.
Most workers are enthusiastic and get used to it very quickly. It is usually quicker to train a new order picker in voice directed picking than paper-based picking.
The speech recognition is not usually affected by a cold, but hoarseness or laryngitis may affect operation. However, voice training can be done on the fly if you experience problems.
As long as the users are not 100% hearing impaired, they can often use the system.
The Talkman terminal has onboard intelligence and can continue to operate out of RF coverage, but the functions available will depend on the application. During order picking the entire order is downloaded into the terminal so that a picker would not normally notice a break in RF coverage. However, in some exceptional circumstances the Talkman will need to communicate with the server, in which case the application waits until the unit is within RF range again.
The picker will normally read back the last 2 or 3 digits of the barcode so that the system can check the correct item has been picked. The picker does not have to read back the entire barcode.
No, speech synthesis and speech recognition is carried out on the Talkman unit, and only data is transmitted over the RF network. If voice were transmitted over the network, the bandwidth would restrict the number of users and high performance voice servers would be required. Because the Talkman system transmits only data and not voice, it does not suffer from these limitations.
There is no practical system limitation to the number of terminals that can simultaneously use the system, as voice is not transmitted over the network. Obviously, the main server must be capable of processing the transactions in real time.
The user's voice template is downloaded to the terminal over the RF network at the start of a shift, and takes only a few seconds.