Mining Contract: Machine Learning Models that Blend Subjective and Objective Data to Improve Understanding and Management of Operator Fatigue
In addition to training a number of masters of science and PhD students in identifying and mitigating mining hazards, this capacity-building contract will take a systematic, comprehensive approach to improve understanding of how to identify and reduce fatigue events in operators, and how to implement a sustainable management program that leads to fewer and less severe incidents and accidents caused by fatigue episodes. The research will apply machine learning to model operator fatigue using production monitoring technology, survey tools, and fatigue monitoring systems.
Contract Status & Impact
This contract is ongoing. For more information on this contract, send a request to email@example.com.
Despite numerous previous interventions and significant advances to improve monitoring of haul truck operators' fatigue levels, effective fatigue management continues to be a persistent issue for mining organizations. Fatigue in mines needs to be better understood, modeled, and systematically managed.
Technology exists that measures potential indications of fatigue (cameras, cognitive tasks, etc.) as they happen; however, these tools do not offer information upon which decision makers can rely to address fatigue at or before the time it occurs in the field. Additionally, as modern mines become increasingly automated (or semi-automated), decreased operator engagement could lead to more fatigue with an even larger impact on operations.
This research will:
- Identify and develop fatigue model parameters
- Model haul truck operator fatigue
- Develop fatigue management instruments
As part of previous work, a large body of operational and fatigue monitoring data has already been acquired from a large mining operation. This data set is one of the largest and most thorough of its kind, and we seek to augment it with additional data sources and expand the date ranges covered.
A critical source of new data will come from a series of focus groups which will be used to develop a survey instrument to assess fatigue related operator issues and the participation of additional mining operations. The goal of this participatory research approach will be to better understand the human factors related to fatigue and how they can be incorporated into a machine learning and analytical model.
Next, machine learning models will be created to predict fatigue in an operation. Through analysis of these models, a set of leading indicators of operator fatigue will be determined. Lastly, a method for linking these survey tools, prediction models, and leading indicators into mine health and safety management systems and processes will be developed.
This research contract will deliver:
- Survey instruments
- Fatigue prediction model for haul truck drivers
- Analysis of leading indicators of fatigue
- Gap analysis of fatigue scheduling algorithm
- Pilot proposal of operationalizing survey and model findings
- Development and Application of Reservoir Models and Artificial Neural Networks for Optimizing Ventilation Air Requirements in Development Mining of Coal Seams
- Modeling and Prediction of Ventilation Methane Emissions of U.S. Longwall Mines Using Supervised Artificial Neural Networks
- Modernization and Further Development of the NIOSH Mine Emergency Response Training System (MERITS), Phase 1
- Probability of Making a Successful Mine Escape While Wearing a Self-Contained Self-Rescuer
- A Simulation Approach Analyzing Random Motion Events Between a Machine and its Operator
- SPONCOM - A Computer Program for the Prediction of the Spontaneous Combustion Potential of an Underground Coal Mine
- SponCom - Spontaneous Combustion Assessment Software - 2.0
- A Study of Burnout in Accident Investigators in the US Mining Industry
- Technology News 545 - NIOSH Updates Spontaneous Combustion Assessment Software
- Technology News 549 - MFIRE 3.0 - NIOSH Brings MFIRE into 21st Century