A human factors analysis of the Stretch mobile manipulator robot

Abstract

Assistive and mobile robots have potential to support everyday domestic tasks and enable independence for persons in the home. As a first step to evaluating this potential, we assessed the initial unboxing and setup of Hello Robot’s Stretch RE1– a novel mobile manipulator designed for domestic settings. All study procedures took place in the McKechnie Family LIFE Home, which is a smart home research facility on the University of Illinois Urbana-Champaign campus. We used subject matter experts (SMEs) and followed human factors principles to consider obstacles users with diverse needs and capabilities (e.g., older adults, persons with mobility disabilities) might encounter during the unboxing process. We then conducted 50 trials of user testing and critical task analyses in the LIFE home to assess the feasibility and usability for different use cases. Research team members controlled Stretch by using a game controller. We used Stretch to manipulate 15 different types of objects that would be part of domestic activities needed to live independently, such as tasks needed for meal preparation. We documented the frequency of errors, time spent manipulating the object, and informal qualitative feedback from teleoperators during and after each trial (using a think-aloud protocol). Implications for future domestic robot design using human factors approaches are discussed.

Development of a Vision-Guided Shared-Control System for Assistive Robotic Manipulators

“…In terms of success rates, from Tables 3 and 7, opening the cabinet was the most challenging subtask for all individuals regardless of their experience with operating a robotic arm. This is similar to the findings by Kadylak et al [30], who found that objects that rotate about a fixed axis were the most difficult to manipulate when teleoperating a mobile manipulator. As a joystick cannot control more than two DoFs at a time, it is difficult to control the end-effector to follow a curved trajectory.…”

“…We described the development and preliminary implementation of a new sharedcontrol approach for ARMs in this paper. This work was built upon our previous work [20,30] and expanded the implementation of VGS control from addressing simple onestep tasks to addressing multi-step functional tasks. We observed that the VGS control delivered similar benefits for multi-step functional tasks as for simple one-step tasks, in…”

Maximizing the Benefits of Participatory Design for Human–Robot Interaction Research With Older Adults

“…Robots are one form of technology being developed to support a range of activities including manufacturing (e.g., Baxter, a multipurpose industrial manufacturing robot; Elprama et al, 2017); search and rescue (e.g., Boston Dynamic’s Atlas robot; Kohlbrecher et al, 2015); delivery services (e.g., Relay, an assistive robot that delivers room-service in hotels; Tussyadiah & Park, 2018); and domestic tasks (e.g., robotic vacuums such as iRobot’s Roomba; Willow Garage’s PR2; Smarr et al, 2014). Increasingly common are assistive robots that can potentially support the daily activities of older adults (reviewed in Wiczorek et al, 2020) or individuals with disabilities (e.g., Hello Robot’s Stretch; Kadylak et al, in press).…”

Are Friendly Robots Trusted More? An Analysis of Robot Sociability and Trust

“…Robot care providers can be purely mobile platforms, mobile manipulators, or completely stationary; range in humanoid features; and have various autonomy capabilities and self-directedness. Specifically, mobile manipulators (e.g., Stretch, Tiago), which are robot caregiver systems with both mobility capabilities and a robotic arm for manipulation [3], can support instrumental activities of daily living (IADLs) such as cleaning, assistance with grabbing, or reaching objects. In contrast, exclusively socially assistive robots do not perform manipulation tasks (e.g., NAO, Jibo), but instead provide social support capabilities for IADLs (e.g., medication management, alarms, and reminders).…”