Process Evaluation

The subject of process evaluations covers a broad range of topics, but the center of the issue is still simple. In each process, the individual must perform a task or a set of tasks. The physical characteristics of the task, as well as the effort required by an individual performing the task, are quantifiable. Methods of quantification are grouped based on the type of data being collected: muscular effort, joint posture, physiological effort, and force application. After quantifying the task requirements and effort of the individual performing the task, results can be used in several ways:

  • To benchmark findings against best practices, standards, or guidelines
  • To recommend areas for improvement
  • To justify recommendations for change

Human-Robot Interaction in Healthcare Tasks


In recent years, there has been a shortage of registered nurses in healthcare facilities. Therefore, nurses are working under high workload and this can lead to errors in providing patient services. Such errors can include misreading a prescription, giving a patient the wrong medicine, or administering incorrect amounts of medicine. To address this issue, robots have been implemented and used in hospitals to assist nurses with basic medicine delivery tasks. However, how patients respond to robots and whether they comply with robot requests to take medication are interesting research questions that need to be investigated. Therefore, the main objective of this study is to apply classic strategies of etiquette in human-human interaction (HHI) to the design of social robot interfaces, and to assess how robot etiquette levels may influence performance of humans and robots, in simulated medicine delivery tasks. Another goal is to better understand the effects of robot physical appearance on user perceptions of robot speech politeness strategies.

An experiment is being conducted in which subjects are instructed to complete Sudoku puzzles while being interrupted by, and interacting with, a medicine delivery robot. (Subjects have been screened for minimum prior experience with Sudoku puzzles.) The subjects are instructed in the puzzle as a primary task and told to complete as many blocks as possible within a 15 minute period. An incentive-based payment system is used in the experiment to motivate the subjects to performance. Subjects are also instructed in advance that a robot will enter the experiment room and attempt to interrupt them and provide them with some medical related-service. The robot employs one of four speech patterns in interacting with subjects, based on Brown and Levisons HHI etiquette model, ranging from a completely direct and impolite strategy to a verbose and profusely polite strategy. The independent variables also include robot physical appearance with two levels, including abstract and human-like. If a subject interacts with the robot, the system provides them with some hints on how to complete the specific Sudoku puzzle (only when finished with the medicine delivery task). The subjects have to decide whether or not to respond to the robot’s request at a specific point in time. All trials are videotaped and end after 15 minutes. Three dependent variables are being recorded:  (1) user subjective perception of robot etiquette in language, (2) user correct completion (and completion rate) of the Sudoku puzzle, and (3) user error rate in the puzzle.

While the study has not yet been completed and findings are not complete, expected outcomes include the following:

  1. Etiquette strategies in HHI can be extended to human-robot interaction in the medicine delivery task and reminding scenario
  2. Users are able to perceive differences in robot language patterns
  3. The robot delivery task will prove to be a distraction and interruption to the subject’s main task and performance
  4. Polite linguistic strategies to conform to social requirements during the service robot delivery task and that these strategies may mediate any adverse effects of robot interruption of user tasks
  5. User performance in the primary task and compliance with robot requests to be greater under high robot etiquette conditions
  6. There will be an interaction effect between robot appearance and robot etiquette level – in terms of user compliance with robot requests; a negative etiquette strategy plus a human-like appearance is expected to be better than all other combinations.

Scaffold Assembly/Disassembly Ergonomics Intervention

Many maintenance and construction tasks result in numerous ergonomic-related injuries. Scaffold assembly / disassembly are two such tasks that present significant ergonomic risks to workers.  Despite such risks, not many studies have been conducted on scaffolding operations. Therefore, a study is being conducted with the nuclear power industry to examine the biomechanical benefits of several ergonomic interventions.

Observation and video-analysis of the scaffolding process has been conducted to help identify specific “High Risk” tasks of interest.  Tasks include (1) ratcheting of clamps and coupling mechanisms for frames, (2) tying down planking with heavy-gauge wire, and (3) handling of planks and scaffold material.  Trained scaffold operators will be recruited for the biomechanical study.  Dependent variables will include: (1) Muscle activity using surface EMG, (2) Postural deviation using electronic goniometers, (3) Displacement, velocity and acceleration at joints using accelerometers, and (4) Spine position, velocity, and acceleration using the Lumbar Motion Monitor.

While the study is in progress, there are no results to date.  However, it is expected that the ergonomic interventions tested will show significant biomechanical benefits in terms of reduced muscle activity, improved posture and joint loading, reduced repetitive movement and forceful exertion required to clamp scaffold joints, and reduced exposure time in awkward postures.  However, the study is being conducted to confirm such hypotheses.

Methods Assessment of Ergonomics Intervention

Industrial assembly, VDT work, and inspection tasks often require awkward, static postures for the neck and shoulder. These postures have been associated with increased musculoskeletal discomfort and decreased productivity. Biomechanically, flexed neck postures increase the load on the neck or cervicobrachial region. Interventions designed to improve worker posture have had the effect of reducing the load on the trapezius muscles and contributed to reductions in sick leave. A funded project investigated the effect of such an ergonomic intervention on a simple assembly task. The intervention took the form of a periscope-style mirror set-up used to reduce the amount of time a worker spends with their neck bent to look at their work.

The project consisted of two studies. In the first study, twelve participants performed a simple, fifteen-minute assembly task driving screws into blocks. The task was performed with four workstation arrangements: (1) industrial standard configuration, (2) mirror configuration, (3) periscope configuration, and (4) pistol grip configuration. In the second study, eight participants performed the same task for four hours in the industrial standard and periscope configurations.

Posture (neck, shoulder, and elbow), muscle activity (bilateral trapezius, cervical and thoracic erector spinae, levator scapulae, and deltoid) productivity, and discomfort data was collected from each subject in each trial condition.

Study 1 results indicated that there were positive biomechanical impacts associated with the interventions. The trade off for these ergonomic improvements was a decrease in productivity during the fifteen-minute trials. Initially, performance is inhibited while using the periscope intervention. It was initially hypothesized that the initial performance decrement would be overcome and the subject would eventually reach the same (or surpass) productivity level when performing the simple task with no intervention. The basis for this hypothesis is grounded in the belief that an individual’s productivity decreases with increased muscle fatigue. In this instance, the muscle fatigue would develop in the muscles in the cervical (neck) spinal region. In study 2, productivity in both conditions improved over the four-hour period. The improvements using the periscope were greater in the periscope configuration compared to the industry standard configuration. Three of the eight participant’s productivity levels at the end of the sessions were the same regardless of the workstation configuration. It is suggested that the productivity decrements associated with the interventions can be mitigated via greater training and adaptation.