An Ontario sawmill has helped develop a voice command lumber inventory management system that offers a quick payback.
By Tony Kryzanowski
An Ontario sawmill is taking a page from retailing giant Wal-Mart by adapting that company’s inventory management system to streamline and reduce costs in managing its own lumber inventory. The Ben Hokum and Son Ltd sawmill is combining the power of wireless communication technology, voice recognition computer software, and data collection computer programs to create a more efficient log scaling, lumber tallying and grading system. A local computer firm, Echo Technology, used the Hokum sawmill as a testing ground and put the system through its paces over the past year.
It is now being marketed to other sawmills, kiln operators, and logging contractors. While the hardware has been readily available for the past decade and is commonly used to manage inventory for businesses such as large retailers and grocery chains, Echo Technology wrote the specific software that makes the system purpose-built for the lumber industry. The system’s benefits include: real time lumber quantity and quality data collection, up-to-the-minute inventory reports, better tracking of inventory shipped to customers, less administration, potentially reduced labour costs, and the ability to hook this electronic inventory control system directly to the sawmill’s accounting system for billing purposes. “I really believe this is something we are going to see in the industry, particularly in Ontario and Quebec, in the next two years,” says Dean Felhaber, president of the Hokum sawmill in Killaloe, Ontario.
“It’s going to save companies a lot of wage and administrative costs.” Realizing that its tallying system was inefficient, the Hokum mill was looking for a better method and was approached by Echo Technology about working together to develop this system. “It was my experience in sawmills —before I became a computer programmer— that inspired me to develop this system,” says Echo Technology president Lawrence Benfield. “I saw an extreme need for streamlining inventory control.” This e-tallying system has three major applications: log scaling, softwood grading and hardwood grading.
However, Felhaber says its potential to improve efficiency is limitless in sawmill applications where tallying of any kind is currently being handled with paper and manual labour for data collection and input. “It’s an electronic means of capturing information,” he says. The Hokum sawmill operates two lines, and manufactures about 36 million board feet of lumber annually, from white pine, red pine, aspen, and dense hardwoods. Of that 36 million board feet, about 25 million is graded.
It wasn’t just improved efficiency that attracted Felhaber to the project. The quick payback on the sawmill’s investment through reduced labour costs was also attractive. Once it was fully implemented, he was able to eliminate seven staff positions. He estimates the overall payback on the investment will take only about 18 months. “Our costs are always increasing and we haven’t seen any increase in the price of our lumber for a number of years,” he says. “Our motivation to provide a testing ground for this new system was to reduce our operating costs, which we’ve done.”
Prior to installation of this system, the Hokum sawmill had a tallyman working with every softwood and hardwood grader. What the system has done is incorporate the tallying and grading function into one position. The grader is now equipped with a data gathering and transmission device about twice the size of a wallet attached to his belt, an earpiece, and a headset with a microphone. While inspecting sticks of lumber travelling along green chains at various locations, each grader voices specific pre-programmed grades, lengths, surface measures and/or widths into the mouthpiece.
These are then transmitted in real time over a wireless communication network into a central computer data collection point. That way, if the grader makes a mistake or the system makes a recognition error, it can be corrected immediately. The system comes complete with its own inventory management system, which automatically adjusts to allow for currently graded lumber and shipments. The production manager has access to up-to-the-minute data on exactly how much lumber has been manufactured to fill specific orders based on the spoken data entered into the system. With pine, the graders specify grade and length with their voice commands.
With hardwood, they specify the length, grade and surface measure. Felhaber says this system is more accurate than the sawmill’s previous method, as each time the grader speaks his command into the microphone, it is repeated back to him through his earpiece. The Hokum sawmill has also installed this wireless data collection and transmission system to handle its log load scaling, although the data collection is not transmitted in real time. Typically, log scaling is managed using a paper tally card and dot system for keeping track of lengths, diameters, time of delivery, and who delivered the logs.
The tally cards were delivered to administrative staff who then entered the information into a computer to arrive at a board foot total. Based on that information, the supplier was paid. Now, the scale operator uses voice commands to input scale information into his data collection waist pack. It can hold data up to an entire day’s delivery. He places his collection unit into a cradle in the scale house, where the information is downloaded into a computer.
The accounting office is connected to the scale house by modem, and can download log delivery information directly into the accounting system. No paper tallying or manual data entry is required—another saving in administrative costs. There were a few challenges to fine-tune the system to ensure that it performed to expectations. The first challenge was to ensure that the voice recognition software would accurately record voice commands spoken into it. Accurate interpretation of speech has been a major complaint related to voice recognition systems currently on the market.
Benfield says the key to this system’s success is that it uses a small vocabulary system, with vocabulary specifically designed for this application. “The fewer words that the recognizer has to try to decipher, the better, and the better the recognition becomes,” he says. Another challenge was to develop a system that could function in a noisy industrial environment.
As with other voice recognition systems, graders need to train the software so that it recognizes their voice and speech patterns. In this case, it was also necessary to filter out sounds such as board clapping, front-end loaders, chainsaws, and other people speaking. A third challenge was setting up the wireless network to ensure that all dead spots in the 130-acre sawmill area were eliminated. Benfield says the objective was to develop a system that is user friendly, and that has been accomplished.
Employees have found using the system no more complicated than speaking into the telephone. “The control software is extremely user friendly,” Benfield says. “We’ve been making computer software for the lumber industry for six years and we know that we are dealing with people who are not regular computer users.” Another friendly feature is that the computer operating system where the data is collected works on a Windows platform. Now that the system has proven itself in actual commercial use, Echo Technology has taken it to market. They have already made presentations to other sawmill owners in Eastern Canada and the northern US states. The response has been positive so far—companies began showing interest in the system as soon as it became available.
This page and all contents
©1996-2007 Logging and Sawmilling
Journal (L&S J) and TimberWest Journal.
last modified on
Tuesday, September 28, 2004