Industrial Speech Recognition

Sinelco-voice-picking-belgiumUnlike consumer speech recognition systems, industrial solutions require near-perfect recognition accuracy, even in a noisy environment with a widely diverse user population. To meet these demands, every Jennifer Mobile application includes the Serenade Speech Recognition Platform. Serenade ensures users are understood quickly and accurately, every time they speak.

Better Recognition Is Critical

Speech recognition accuracy is critical to user adoption and efficiency. If a warehouse worker using a voice picking system has to repeat him- or herself, or if the speech recognizer mis-recognizes commands, the user’s productivity suffers. Equally important, recognition errors frustrate users and can lead to worker resistance to using the technology.

From a dollars and cents perspective, every five percent difference in recognition accuracy – 99% accuracy versus 94%, for example – translates into 2-6 minutes of productivity benefit per user per day (depending on pick rates and other factors). While the individual time benefit of better accuracy is small, the cumulative benefit for operations with large numbers of users is significant.

Technology Overview

Serenade incorporates the latest advances in speech recognition technology with additional noise filtering and audio processing technologies to ensure near-perfect recognition in harsh industrial settings. Read our white paper for a full description of speech technology used in a DC.

Dual Recognition Technology

Serenade utilizes multiple simultaneous recognition strategies (phonetic and word-based) to ensure optimal recognition, even for users with non-standard accents or non-standard speech patterns who may struggle to get acceptable recognition with other speech systems.

No Training

Unlike consumer-oriented recognition systems used for automated call routing and other customer service applications, industrial speech systems are designed to optimize recognition for each individual user. While traditional speaker-dependent systems require 20-30 minute voice training sessions, Serenade requires only a 5-minute user enrollment process. (While people may start using Serenade without enrolling, the system is not speaker-independent. Read our White Paper to learn why this is an important distinction.)

Ongoing Adaptation

Serenade automatically adapts to sudden and frequent changes in background noise, and the system adapts to users, even as their speech patterns change in the course of a single shift (due to fatigue, for example), or from day-to-day. With many legacy voice recognition systems, users frequently had to record a new model as they became accustomed to the system and started to speak more quickly. With Serenade users never need to retrain, and they are able to tune the system to better recognition individual words or phrases.

Download our white paper, The Evolution of Speech Recognition Technology In the Warehouse for a more in-depth introduction.