Vision-based Ascending Staircase Detection with Interpretable Classification Model for Stair Climbing Robots

Lee, Kangneoung; Kalyanram, Vishnu; Zhengl, Chuanqi; Sane, Siddharth

doi:10.1109/icra46639.2022.9812456

Cited by 6 publications

(8 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, to utilize the advanced locomotion capabilities of α-WaLTR, new algorithms must be added and modifications to existing ones are necessary. The newly developed algorithms include stair detection [64], visualbased stair navigation, stair climbing, and velocity arbitrator algorithms. The yellow boxes in bold text in Fig.…”

Section: A Ros-based System Architecturementioning

confidence: 99%

“…This subsection presents an efficient stair detection strategy that can be applied for embedded robotic systems. While many machine-learning-based algorithms exist [65], [66], [67], [68], [69], [70], most of them are not directly suitable for embedded robotic applications due to 1) models trained using a significantly different dataset, resulting in low performance when implemented in the robot [64]; and 2) high computing and processing load for real-time operations and retraining when needed [64], [70].…”

Section: B Stair Detection Algorithmmentioning

confidence: 99%

“…We have recently developed a stair detection algorithm, which allows fast and reliable stair detection using RGB-D data [64]. In field operations, the environment can be significantly different from the data used in establishing the classification model.…”

Section: B Stair Detection Algorithmmentioning

confidence: 99%

“…The three features representing the characteristic of a staircase are inputted to the interpretable classification model to determine whether the image has a staircase or not. The evaluation performed in comparisons with an SVM-based method [68] and You Look Only Once [70] show that the adopted method is comparable or better in stair detection performance while requiring less computational time and cost [64].…”

Section: B Stair Detection Algorithmmentioning

confidence: 99%

See 3 more Smart Citations

$\mathbf{\alpha }$-WaLTR: Adaptive Wheel-and-Leg Transformable Robot for Versatile Multiterrain Locomotion

et al. 2023

Self Cite

View full text Add to dashboard Cite

Adaptability is a fundamental yet challenging requirement for mobile robot locomotion. This article presents α-WaLTR, a new adaptive wheel-and-leg transformable robot for versatile multiterrain mobility. The robot has four passively transformable wheels, where each wheel consists of a central gear and multiple leg segments with embedded spring suspension for vibration reduction. These wheels enable the robot to traverse various terrains, obstacles, and stairs while retaining the simplicity in primary control and operation principles of conventional wheeled robots. The chassis dimensions and the center of gravity location were determined by a multiobjective design optimization process aimed at minimizing the weight and maximizing the robot's pitch angle for obstacle climbing. Unity-based simulations guided the selection of the design variables associated with the transformable wheels. Following the design process, α-WaLTR with an embedded sensing and control system was developed. Experiments showed that the spring suspension on the wheels effectively reduced the vibrations when operated in the legged mode and verified that the robot's versatile locomotion capabilities were highly consistent with the simulations. The system-level integration with an embedded control system was demonstrated via autonomous stair detection, navigation, and climbing capabilities.

show abstract

Section: A Ros-based System Architecturementioning

confidence: 99%

Section: B Stair Detection Algorithmmentioning

confidence: 99%

Section: B Stair Detection Algorithmmentioning

confidence: 99%

Section: B Stair Detection Algorithmmentioning

confidence: 99%

See 2 more Smart Citations

$\mathbf{\alpha }$-WaLTR: Adaptive Wheel-and-Leg Transformable Robot for Versatile Multiterrain Locomotion

et al. 2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…Identifying a staircase to prepare for subsequent climbing motion is one of the fundamental challenges for the control of mobile robots and unmanned ground vehicles [15]. Lower-limb prostheses and assistive devices use different actuator control algorithms in different walking conditions, such as walking, running, stair climbing, and so on, to provide the most convenience to the user [16,17].…”

Section: Staircase Detectionmentioning

confidence: 99%

Vision-Based Recognition of Human Motion Intent during Staircase Approaching

Islam

Haque

Imtiaz

et al. 2023

Sensors

View full text Add to dashboard Cite

Walking in real-world environments involves constant decision-making, e.g., when approaching a staircase, an individual decides whether to engage (climbing the stairs) or avoid. For the control of assistive robots (e.g., robotic lower-limb prostheses), recognizing such motion intent is an important but challenging task, primarily due to the lack of available information. This paper presents a novel vision-based method to recognize an individual’s motion intent when approaching a staircase before the potential transition of motion mode (walking to stair climbing) occurs. Leveraging the egocentric images from a head-mounted camera, the authors trained a YOLOv5 object detection model to detect staircases. Subsequently, an AdaBoost and gradient boost (GB) classifier was developed to recognize the individual’s intention of engaging or avoiding the upcoming stairway. This novel method has been demonstrated to provide reliable (97.69%) recognition at least 2 steps before the potential mode transition, which is expected to provide ample time for the controller mode transition in an assistive robot in real-world use.

show abstract

Understanding the Influence of Context on Real-World Walking Energetics

Baroudi

Barton

Cain

et al. 2023

Preprint

View full text Add to dashboard Cite

Speeds that minimize energetic cost during steady-state walking have been observed during lab-based investigations of walking biomechanics and energetics. However, in real-world scenarios, humans walk in various contexts that elicit different walking strategies, which may not always prioritize minimizing energetic cost. To investigate whether individuals tend to select energetically optimal speeds in real-world situations and how contextual factors influence gait, we conducted a study combining data from lab and real-world experiments. Walking kinematics and context were measured during daily life over a week (N = 17) using wearable sensors and a mobile phone. To determine context, we utilized self-reported activity logs, GPS data, and follow-up exit interviews. Additionally, we estimated energetic cost using respirometry over a range of gait speeds in the lab. Cost of transport during these trials was used to identify an energetically optimal walking speed range for each participant. The proportion of real world steady-state stride speeds within this range was identified for all data and for each context. We found that participants walked significantly faster than what was energetically optimal, and the proportion of steady-state strides within the energetically optimal speed range was dependent on the context. On average, 45.6% of all steady-state stride speeds were energetically optimal across all contexts for all participants. These results suggest that while energetic cost is a factor considered by humans when selecting gait speed in daily life, it is not the sole determining factor. Context contributes to the observed variability in movement parameters both within and between individuals.

show abstract

Vision-based Ascending Staircase Detection with Interpretable Classification Model for Stair Climbing Robots

Cited by 6 publications

References 24 publications

$\mathbf{\alpha }$-WaLTR: Adaptive Wheel-and-Leg Transformable Robot for Versatile Multiterrain Locomotion

$\mathbf{\alpha }$-WaLTR: Adaptive Wheel-and-Leg Transformable Robot for Versatile Multiterrain Locomotion

Vision-Based Recognition of Human Motion Intent during Staircase Approaching

Understanding the Influence of Context on Real-World Walking Energetics

Contact Info

Product

Resources

About