Kuo Chen

Rider Trunk and Bicycle Pose Estimation With Fusion of Force/Inertial Sensors

Estimation of human pose in physical human-machine interactions such as bicycling is challenging because of highly-dimensional human motion and lack of inexpensive, effective motion sensors. In this paper, we present a computational scheme to estimate both the rider trunk pose and the bicycle roll angle using only inertial and force sensors. The estimation scheme is built on a rider-bicycle dynamic model and the fusion of the wearable inertial sensors and the bicycle force sensors. We take advantages of the attractive properties of the robust force measurements and the motion-sensitive inertial measurements. The rider-bicycle dynamic model provides the underlying relationship between the force and the inertial measurements. The extended Kalman filter-based sensor fusion design fully incorporates the dynamic effects of the force measurements. The performance of the estimation scheme is demonstrated through extensive indoor and outdoor riding experiments.

Whole-Body Pose Estimation in Human Bicycle Riding Using a Small Set of Wearable Sensors

Tracking whole-body human pose in physical human-machine interactions is challenging because of highly dimensional human motions and lack of inexpensive, nonintrusive motion sensors in outdoor environment. In this paper, we present a computational scheme to estimate the human whole-body pose with application to bicycle riding using a small set of wearable sensors. The estimation scheme is built on the fusion of gyroscopes, accelerometers, force sensors, and physical rider-bicycle interaction constraints through an extended Kalman filter design. The use of physical rider-bicycle interaction constraints helps not only eliminate the integration drifts of inertial sensor measurements but also reduce the number of the needed wearable sensors for pose estimation. For each set of the upper and the lower limb, only one tri-axial gyroscope is needed to accurately obtain the 3-D pose information. The drift-free, reliable estimation performance is demonstrated through both indoor and outdoor riding experiments.

A robotic bipedal model for human walking with slips

Slip is the major cause of falls in human locomotion. We present a new bipedal modeling approach to capture and predict human walking locomotion with slips. Compared with the existing bipedal models, the proposed slip walking model includes the human foot rolling effects, the existence of the double-stance gait and active ankle joints. One of the major developments is the relaxation of the nonslip assumption that is used in the existing bipedal models. We conduct extensive experiments to optimize the gait profile parameters and to validate the proposed walking model with slips. The experimental results demonstrate that the model successfully predicts the human recovery gaits with slips.

Pose estimation in physical human-machine interactions with application to bicycle riding

Tracking whole-body human pose in physical human-machine interactions such as bicycling is challenging because of highly-dimensional human motions and lack of inexpensive, effective motion sensors in outdoor environment. In this paper, we present a computational scheme to estimate the whole-body pose in human-machine interaction with application to the rider-bicycle system. The estimation scheme is built on the fusions of gyroscopes, accelerometers and force sensors with six Extended Kalman filter designs. The use of physical human-machine interaction constraints further helps to eliminate the integration drifts of inertial sensors measurements and also to reduce the number of the inertial sensors for whole-body pose estimation. For each set of upper- and lower-limb, only one tri-axial gyroscope is needed to accurately obtain the pose information. The performance of the drift-free, reliable estimation scheme is demonstrated through both the indoor and outdoor bicycle riding experiments. The proposed approach can be further extended to other types of physical human-machine interactions.

Modeling of rider-bicycle interactions with learned dynamics on constrained embedding manifolds

Modeling and control of physical human-machine interactions (pHMI) are challenging due to the high-dimensional movement of human body. In this paper, we present a hybrid statistical/physical dynamic model scheme to capture the pHMI through a rider-bicycle interaction example. We use the Gaussian process dynamical model (GPDM) to capture the high-dimensional human movement into a low-dimensional latent space. We extend the GPDM by incorporating additional physical control inputs into the model. The GPDM control inputs are coupled with the physical dynamic model from the bicycle systems such as crank angle etc. The proposed statistical/physical dynamic model is further enhanced by constrained manifold learning algorithms so that we can use less training data sets to obtain the more accurate model. We illustrate the modeling scheme through a lower-limb pedaling example in human bicycling experiments.

An integrated physical-learning model of physical human-robot interactions with application to pose estimation in bikebot riding:

Modeling physical human-robot interactions (pHRI) is important in studying human sensorimotor skills and designing human assistive and rehabilitation systems. One of the main challenges for modeling pHRI is the high dimensionality and complexity of human motion and its interactions with robots and the environment. We present an integrated physical-learning pHRI modeling framework with applications to the bikebot riding example. The modeling framework contains an integrated machine-learning-based model for high-dimensional limb motion with a physical-principle-based dynamic model for the human trunk and an interacted bicycle-like robot (bikebot). A new axial linear embedding algorithm is used to obtain the low-dimensional latent dynamics for highly redundant human limb movement. The integrated physical-learning model is then used to estimate human motion through an extended Kalman filter design without using any sensors attached to the limb segments. Extensive bikebot riding experiments are conducted to validate and demonstrate the integrated pHRI model. Comparison results with other machine-learning-based models are also presented to demonstrate the superior performance of the proposed modeling framework for bikebot riding.

An Integrated Physical-Learning Model of Physical Human-Robot Interactions: A Bikebot Riding Example

One of the main challenges for modeling human-robot interactions (pHRI) is the high dimensionality and complexity of human motion. We present an integrated physical-learning modeling framework for pHRI with applications on the bikebot riding example. The modeling framework contains a machine learning-based model of high-dimensional limb motion coupled with a physical principle-based dynamic model for the human trunk and the interacted robot. A new axial linear embedding (ALE) algorithm is introduced to obtain the lower-dimensional latent dynamics for redundant human limb motion. The integrated physical-learning model is used to estimate the human motion through an extended Kalman filter design. No sensors are required and attached on human subject’s limb segments. Extensive bikebot riding experiments are conducted to validate the integrated human motion model. Comparison results with other machine learning-based models are also presented to demonstrate the superior performance of the proposed modeling framework.© 2014 ASME

Dynamic rider/bicycle pose estimation with force/IMU measurements

Pose estimation of human motor skills such as bicycling in natural environments is challenging because of highly-dimensional human motion. In this paper, we present a dynamic rider/bicycle pose estimation scheme that can be used in outdoor environments. The proposed estimation scheme is based on the integration of the rider/bicycle dynamic model with the measurements from the force sensor and inertial measurement units (IMU). We take advantages of the attractive properties of both the force and IMU sensors in the design, that is, the force measurements do not suffer drifting while the IMU measurements generate real-time attitude and acceleration information. The rider/bicycle dynamic model provides an underlying relationship between the force and the IMU measurements. We demonstrate the effectiveness and performance of the pose estimation design through extensive bicycle riding experiments.

Balance recovery control of human walking with foot slip

We present a balance recovery control design for human walking with foot slip. The control strategy is built on the two-mass linear inverted pendulum model (LIP) that represents the human body and limb motions. We first validate the model through experiments of human normal walking and walking with foot slip. We then design a balance recovery control using the capture point (CP) concept. We extend the CP-based walking control and incorporate time-varying locations of the zero moment point. These extensions allow the balance recovery control of the humans center of the mass movement to rapidly respond to the unexpected foot slip. We conduct experiments to tune the model parameters and to validate the slip recovery control.

Slip detection and prediction in human walking using only wearable inertial measurement units (IMUs)

Slip and fall is one of the major causes for human injuries for elders and professional workers. Real-time detection and prediction of the foot slip is critical for developing effective assistive and rehabilitation devices to prevent falls and train balance disorder patients. This paper presents a novel real-time slip detection and prediction scheme with wearable inertial measurement units (IMUs). The slip-detection algorithm is built on a new dynamic model for bipedal walking with slips. An extended Kalman filter is designed to reliably predict the foot slip displacement using the wearable IMU measurements and kinematic constraints. The proposed slip detection and prediction scheme has been demonstrated by extensive experiments.