This page presents some highlights of current and recent projects. See the Research page for a broader overview of my research activities.
I gave a talk at the Stapp Car Crash Conference in Orlando this month. My paper examined the effects of driver characteristics on seat belt fit. We'd previously shown that obesity significantly worsens lap belt fit in passenger conditions. The latest study, based on data from 97 men and women with a wide range of age and body size, show that many drivers place their seat belts farther from their pelves than they should. Individuals with higher BMI had worse belt fit, on average. The results suggest a need to educate drivers about proper belt placement, but also indicate that improved belt designs might help. This research was funded by the Toyota Collaborative Safety Research Center.
I gave a short talk at the IRCOBI conference in Gothenburg Sweden last month. My paper focused on our work measuring and modeling seated occupant body shapes for parametric human body modeling. We are continuing to improve the depth and breadth of our human modeling capability. We have a rich dataset that includes children ages 2 and up, young adults, elderly, and soldiers.
I participated in an informative symposium hosted by SAFER in Gothenburg Sweden. With a focus on occupant protection for children 4-12 years old, the symposium noted the considerable progress made in the past decade while outlining the remaining challenges. The presentations are available at this website.
My colleague Matt Parkinson and I hosted a symposium on digital human modeling in Ann Arbor. We had over 140 participants from 12 countries with poster and podium sessions over three days. The program, proceedings, and other info are available at dhm2013.org. One highlight was a twitter scavenger hunt featuring our brand-new human modeling avatar, Diablos.
My colleagues and I presented an update on our work on soldier posture, position, and body shape at the 19th annual meeting of the Automotive Research Center at the University of Michigan. The overall goal of this work is to develop modeling and simulation tools that will lead to safer and more efficient vehicle designs.
Driver preference for fore-aft steering wheel position has not previously been addressed directly in the SAE literature. We analyzed data from a laboratory study to create a logistic regression model predicting subjective response as a function of H30, L6, and driver stature. The results compare favorably to a large set of vehicle benchmarking data.
My colleague Jingwen Hu developed a new framework for conducting virtual seat fit testing using a parametric human body model. This work differs from previous simulation studies of human/seat interaction in using a human body model that can represent a large range of male and female body shapes and postures. Our ongoing body-shape modeling work includes the development of new, whole-body paramtric shape and posture models based on a wider range of postures than are found in the CAESAR database we used for this study.
We have measured the interior geometry of dozens of vehicles over the past few years for a wide range of safety and ergonomics applications. One important finding for safety is that the seat belt anchorage locations in second-row, outboard seats vary widely, with the lab belt anchorages spanning essentially the entire legal range. ATD and simulator studies show that the range of belt fit we document in our current paper can have important safety implications.
We've just completed development of anthropometric specifications for the Army's new WIAMan blast dummy. The new anthropomorphic test device (ATD) will lead to the development of safer vehicles, seats, and interiors. This work is an outgrowth of our Seated Soldier Study.
The U.S. Army's TARDEC organization recently featured our work in an internal publication. The data from this study will have broad applicability in the development of road vehicles, not just military vehicles. This study is the first to gather and model data on whole-body shape from a large, diverse sample of people in supported seated postures. Consistency in study protocol will allow the data from the 300 soldiers measured in this project to be merged with data from 200 adults measured in another current project.
Recent research, including studies by my colleagues at UMTRI, has shown that older drivers are at increased risk of injury in vehicle crashes. Some of that risk is due to frailty -- injury (particularly thoracic injury) occuring with less loading than is required to injury younger occupants. But the posture and belt fit of older drivers may also put them at greater risk. To addess this issue, we are conducting a laboratory study of older occupant posture, belt fit, and body shape. Supported by the Toyota Collaborative Safety Research Center, the study measured 200 men and women, including 120 over age 60. The participants sat in laboratory mockups that can be reconfigured to represent the driver and passenger package configurations of a wide range of vehicles, from sports cars to minivans. Posture and belt fit are recorded using a FARO Arm coordinate digitizer to measure body landmark locations. A whole-body laser scanner is used to record the participant's body shape in a range of postures. These data will be valuable for designing improved restraint systems that optimally protect occupants of all ages.
At the 2012 Human Factors and Ergonomics Society Annual Meeting in Boston, I presented an overview of the methods for our recent large-scale child anthropometry study. We used a laser scanner to record the body shapes of 162 kids age 4 to 11 in a wide range of standing and seated postures. These data will be used to develop new anthropometric specifications for crash dummies, but also have application to a wide range of other products for children. We will be using the data to develop parametric finite-element models of children for use in crash simulation and restraint-system optimization.
Wired Autopia covered some of our recent work on body shape modeling. We're using a range of technologies to measure posture and body shape in automotive postures. Although the results have significant applications in ergonomics (for example, improved seat design for comfort) the primary applications of the data and resulting models will be in the development of improved physical and computational surrogates for crash simulations. These tools will allow improved optimization of restraint systems to proteect vehicle occupants with a wide range of physical characteristics. This presentation covers a some of our current activities.
We collaborated with Anthrotech on a large-scale study of soldier anthropometry, focusing on seated postures and body shapes with a range of clothing and personal protective equipment. The goal of the research is to develop new quantitative models that can be used for the design of physical and computational surrogates used for designing and assessing military vehicles. One near-term application is the Warrior Injury Assessment Manikin (WIAMan), a new anthropomorphic test device (ATD) to be used for evaluating protection in underbody blast events. UMTRI will use the data from the seated soldier study to specify the posture and body shape for WIAMan. Other applications include the development of statistical models to predict posture for soldiers in both driver and crew workstations. These models will be used to position computational models of soldiers for both ergonomics analysis and blast simulations.
The prevalence of obesity among US adults has plateaued over the past few years, but still remains at historical highs. Analysis of crash data has suggested that, after correcting for other factors, obesity increases the risk of certain types of injuries. One reason for that increased risk may be problems with seat belt fit. We conducted a laboratory study of belt fit among 54 adults, 48% of whom were obese. We found that obesity results in relatively poor lap belt fit, with the belt riding high and forward relative to the pelvis. Both lap belt and shoulder belt length were also increased for heavier individuals, a trend which also reduces belt effectiveness. These results indicate that load sharing among the components of restraint system (including the belt, airbag, and knee bolster) may differ substantially for obese adults and should be taken into account in restraint system design.
We have added new content to trucksteps.org, our website devoted to safe ingress and egress for truck drivers. Browse on over and check it out!
We are conducting a major study of child posture and body shape. This is the first study to gather three-dimensional data on child posture in automotive seated postures. We have measured 160 children spanning an age range from45 to 11 years seated in a rear seat from a sedan, with and without a belt positioning booster, and in up to 18 different postures in our 3D body scanner. The primary near-term application of thedata will be to improve the design of crash dummies and computational models of child occupants. But the data will also have broad application for the design of child restraints, protective equipment, clothing, and other products and environments that interact physically with children in this age range.
Digital human modeling for ergonomics applications is the focus of my research in the Center for Ergonomics. The research I direct in the Human Motion Simulation Laboratory focuses on the development of algorithms and models for accurate prediction of human postures and motions in tasksituations relevant to product and workspace design. We emphasize algorithms that can be implemented in any human figure model, rather than limiting ourselves to the capabilities of one model. We are fortunate to have the developers of the two most popular software systems for industrial ergonomics (Dassault Systemes and Siemens PLM) engaged in our laboratory as technology partners. Among other topics, we're currently working on improving grasp simulation. The literature on grasp is vast, including contributions from motor control and robotics as well as ergonomics and human modeling. Yet, ergonomics using human models spend a large amount of time working to simulate grasp, often with poor results. As part of ongoing research in the Human Motion Simulation Lab, Ph.D. candidate Wei Zhou and I developed a data-based method for predicting grasp motions. The hand motions are parameterized on target size and grasp type, so that they're readily configurable for different objects. We integrated this kinematic prediction with collision sensing to enable automated grasp of objects with arbitrary geometry.
Three-dimensional anthropometry is a part of many of my current research projects. Using data from UMTRI's whole-body laser scanner, medical imaging data, or studies conducted elsewhere, we create statistical models of anatomical shapes. These models are used for a wide variety of design and analysis purposes, from creating anthropometric specifications for crash dummies to virtual seat assessments. One challenge in using 3D scan data is that each scan represents a single posture. For general applications, it's necessary to be able to alter the posture of the scan. This figure shows a statistical model of the seated torso that encompasses both anthropometric and postural variability. A single relaxed seated scan from each of 712 men and women was morphed to 15 different postures using a skeletal linkage. The resulting external body shapes were analyzed using a principal component analysis/regression (PCAR) approach to produce the model demonstrated in the figure. The model provides an efficient way of representing a wide range of seated body shapes driven by both structural anthropometry and posture.The model has applications from seat and chair design to the development of parametric finite-element models for crash safety analysis.
The design of truck and off-highway equipment cabs (agriculture, construction, mining, and forestry) is becoming more complex due to the increase in the use of electronic controls and displays. In most cases, the new equipment is added to the customary manual controls, creating a need to optimize the layout for all tasks.
We conducted a study of seated operators to quantify reach and force-exertion capability as a function of task location relative to the driver and package. Licensed truck drivers reached for push-button targets located throughout the workspace, providing subjective ratings of difficulty for each target. The subjects also exerted force on a handle located in 13 discrete positions spanning the range of typical hand control locations.
The drivers also operated accelerator, brake, and clutch pedals, using both normal motion speed and "as fast as possible", simulating emergency operation. An optical motion-capture system was used to record whole-body kinematics.
The data from this study are being used to develop new design tools for assessing and laying out truck and heavy equipment cabs.
Truck driving is among the most dangerous occupations in the U.S. Crashes are not the main cause of injury, though: Most occupational injuries to truck drivers are due to slips and falls. With support from at R01 grant from the National Institute for Occupational Safety and Health, we are conducting a field and laboratory study of truck driver ingress and egress. Our work uses biomechanical analysis techniques to quantify ingress/egress behaviors. We have gathered field data on entry and exit techniques drivers use, interviewed drivers on their experiences, and conducted a detailed laboratory study. Using an optical motion capture system and an instrumented cab mockup, we recorded driver's movements and the forces they exerted with a wide range of step and handhold configurations.
The outcomes of this research will include new design guidelines for truck ingress/egress systems (steps and handholds), new biomechanical analysis methodologies that can be used to assess candidate systems, and a better understanding of the factors that influence the risk to drivers of both acute and chronic injury.
The performance of belt-positioning booster seats used by children who are too small to achieve good belt fit with vehicle belts alone, and no longer riding in harness restraints, is assessed in U.S. regulatory tests using Hybrid-III dummies. Unfortunately, these dummies interact with belt restraints in unrealistic ways. The lap, shoulder, and chest areas where the belt contacts the dummy are of particular concern, because the dummy geometry in these areas is substantial different from the skeletal anatomy of similar-size children. We are working on developing improved components for the dummies, starting with the Hybrid-III six-year-old. We have developed new methods for extracting and analyzing the shapes of anatomical structures from medical imaging data. We've used these methods to create a statistical model of the child pelvis, thorax, and shoulder based on data from over 80 children from ages 4 to 12. The analysis allows us to create skeletal models for children in that age range as a function of age, stature, body weight, or other variables. We've used these model to create a geometric target for the 6YO pelvis and thorax. This figure compares the skeletal gometry of the Hybrid-III 6YO to the targets.
The National Highway Traffic Safety Administration (NHTSA) has amended Federal Motor Vehicle Safety Standard (FMVSS) 202 to require higher and more-forward head restraints (sometimes called head rests, but primarily intended to protect occupant's necks in rear collisions). The new regulation, called FMVSS 202a, requires that head restraints lie within 55 mm of a Head Restraint Measurement Device representing a midsize-male head shape when measured at a seat back angle of 25 degrees. When this rule was proposed, my colleagues and I published a report showing that most drivers sit with more-upright seatback angles. We concluded that designing to the 55-mm "backset" would cause the head restraint to interfere with the preferred head locations of a substantial number of drivers. This prediction has since been substantiated by other research and field reports from vehicles built to the new standard. Recently, my colleague Prof. Matt Parkinson, who heads the OPEN Design Laboratory at Pensylvania State University, conducted a simulation study of an alternative seat back design that optimizes the head restraint position across seat back angles. This approach reduces the likelihood of interference with the head restraint for smaller-stature drivers while ensuring that taller drivers, who tend to select more-reclined seat back angles, are equally well protected. More...
©2017 Matthew P. Reed and The University of Michigan