Click on “Download PDF” for the PDF version or on the title for the HTML version.
If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.
Categorization and Analysis of Past Farm Fatality Incidents to Inform Autonomous Machine Design and Risk Analysis
Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.org
Citation: 2021 ASABE Annual International Virtual Meeting 2101006.(doi:10.13031/aim.202101006)
Authors: Kelly Sandner, John Shutske, Luke Powers
Keywords: Agriculture, Autonomous, CDC, FACE, Hazard Analysis, NIOSH, Risk Assessment, Safety, Tractors
Abstract. Despite decades of technology and machinery advancements in agriculture, farmers, ranchers, and agricultural managers still had a fatal injury rate of 23.2 deaths per 100,000 full-time equivalent workers in 2019 based on rates published by the U.S. Bureau of Labor Statistics (BLS). Now, as fully autonomous machinery is being developed with widespread deployment expected soon in fields and on farmsteads, understanding the circumstances, conditions, and situations that have contributed to past machine-related farm fatalities can contribute to risk assessment efforts. Better risk assessment and control can improve the safety of autonomous machinery, particularly if done in the early phases of design. For designs as novel as autonomous agricultural machines, the lack of historical data limits the efficacy of traditional risk assessment tools such as failure modes and effects analysis (FMEA). Analyzing historic hazards and resultant injuries and deaths can reveal patterns used to inform risk assessment techniques embedded into new and existing engineering standards that would be better suited for autonomous machinery. This study analyzed 434 publicly available fatality records for deaths that occurred from 1990 to 2020. Records had been collected and analyzed by the National Institute for Occupational Safety and Health (NIOSH) through their Fatality Assessment and Control Evaluation (FACE) Program. Most original NIOSH FACE investigations included detailed data collection, often through on-site investigation activities. The scope of this study was limited to tractor-related incidents only. Based on documented FACE report data, each incident was categorized as having been influenced by one or more of: human error, machine error/malfunction, environmental factors, rollover protection structures (ROPS), sloping terrain, and user inexperience. Two reviewers classified each incident independently using a specific methodology to reduce researcher bias. Final categorization was based on consensus of the reviewers. A total of 52% of deaths involving a tractor occurred on sloping terrain. Within analyzed reports, 53% of incidents included a recommendation to install rollover protection structures. Among cases classified as “machine error,” 93% of those incidents involved an experienced operator. Across all reports analyzed, the majority of tractor incidents occurred in situations that were cited as being outside of ideal operating conditions. This data has shown that when humans encounter certain obstacles, or combinations of hazards, there is a pattern of behavioral responses that results in operator injury or machine damage. These responses are the result of limited sensory capability to detect hazards, improper response time, and/or improper decision-making during the hazard avoidance stage. In order for autonomous vehicles to avoid the same instances occurring, they must either have faster hazard detection processing or the ability to detect and process more hazardous conditions than human operators in order to make a safe maneuver.
(Download PDF) (Export to EndNotes)