AI Case Study
BBMV reduced construction worker fatigue and impairment time while on duty by 24% with sleep monitoring devices
BBMV has construction workers wearing sleep monitoring devices which feed data through fatigue detecting algorithms into an app which allows the workers and their supervisors to gauge their level of alertness and predict problems further in their shifts.
Construction And Engineering
"In autumn 2016, BBMV became one of the first organizations in the world to employ a Predictive Fatigue Monitoring program in their operation. While Fatigue Science has provided fatigue measurement tools to industrial firms, the US military and elite sporting teams for close to a decade, the launch of their new Predictive Fatigue Monitoring Solution is now rapidly changing the game in industrial health and safety. The program combines wearable technology with SAFTE, a US Army-developed algorithm, to measure and predict the fatigue level of BBMV workers as they arrive for duty. Using Fatigue Science’s algorithms in advanced simulations of shift schedules, BBMV was able to implement a scheduling change that has reduced risk even further."
"Given that 1) the connection between sleep/circadian factors and fatigue has been well-established, that 2) actigraphy in general has been validated as an acceptable method for measuring sleep/wake patterns and sleep quality in normal healthy individuals and in people with certain types of sleep problems, and that 3) the SAFTE model has been validated as an accurate predictor of performance effectiveness and fatigue risk provided that accurate historical sleep/wake information is available, the only remaining issue is the degree to which a specific actigraph and its associated data-processing algorithms are capable of accurately reflecting the actual sleep/wake patterns of typical individuals."
Health And Safety
"Fatigue has remained one of the largest systemic problems in global health and safety across a variety of heavy industries — particularly in mining, construction, and transportation."
One of the shift supervisors claims that "the Predictive Fatigue Monitoring program empowers him to obtain predictive, actionable information every day, helping BBMV proactively address the most extreme fatigue risks that might arise later during that shift. Now, that information is not only knowable, it’s available hours before that fatigue creates a visible threat to safety. In the first two months of exposing workers to their own fatigue via the Fatigue Science mobile app, BBMV has already seen a 24% reduction in time spent fatigue impaired on duty, and expects to see that number continue to improve as health staff support workers in addressing the root causes of fatigue based on real data."
A proprietary algorithm is used to judge if the wearer of the Readiband actigraph has meet the criteria for sleeping or being awake. This algorithm uses a "combination of smoothing, thresholding, and dilation procedures" which are AI-related techniques.
"The Readiband captures high-resolution sleep data, validated against the clinical gold standard of polysomnography with 92% accuracy. Sleep data is transmitted to the cloud automatically for SAFTE Fatigue Model analysis as soon as workers arrive for duty."