Neural network balance control of hopping robots in flight phase under unknown dynamics

2015 ◽  
Author(s):  
Hicham Chaoui ◽  
Suruz Miah ◽  
Mohamed Redouane Kafi ◽  
Bekhada Hamane
Keyword(s):  
Neural Network ◽  
Balance Control ◽  
Flight Phase ◽  
Hopping Robots

Neural network estimation of balance control during locomotion

2005 ◽  
Vol 38 (4) ◽  
pp. 717-724 ◽  
Author(s):  
Michael E. Hahn ◽  
Arthur M. Farley ◽  
Victor Lin ◽  
Li-Shan Chou
Keyword(s):  
Neural Network ◽  
Balance Control ◽  
Network Estimation

Neural Network Models for Estimation of Balance Control, Detection of Imbalance, and Estimation of Falls Risk

2011 ◽  
pp. 217-242
Author(s):  
Michael E. Hahn ◽  
Arthur M. Farley ◽  
Li-Shan Chou
Keyword(s):  
Neural Network ◽  
Balance Control ◽  
Current Model ◽  
Falls Prevention ◽  
Body Motion ◽  
Estimation Model ◽  
Neural Network Models ◽  
Balance Impairment ◽  
Falls Risk ◽  
Artificial Neural Network Ann

Gait patterns of the elderly are often adjusted to accommodate for reduced function in the balance control system. Recent work has demonstrated the effectiveness of artificial neural network (ANN) modeling in mapping gait measurements onto descriptions of whole body motion during locomotion. Accurate risk assessment is necessary for reducing incidence of falls. Further development of the balance estimation model has been used to test the feasibility of detecting balance impairment using tasks of sample categorization and falls risk estimation. Model design included an ANN and a statistical discrimination method. Sample categorization results reached accuracy of 0.89. Relative risk was frequently assessed at high or very high risk for experiencing falls in a sample of balance impaired older adults. The current model shows potential for detecting balance impairment and estimating falls risk, thereby indicating the need for referral for falls prevention intervention.

Neural network-based optimization model for sprint data collection

2021 ◽  
pp. 1-11
Author(s):  
Guangfei Luo
Keyword(s):  
Neural Network ◽  
Data Collection ◽  
Data Acquisition ◽  
Bp Neural Network ◽  
Optimization Model ◽  
Balance Control ◽  
Performance Parameters ◽  
Advantages And Disadvantages ◽  
Network Algorithm ◽  
Neural Network Algorithm

Sprint data has the characteristics of quality and continuity, but due to the limitations of optimization algorithm, the existing sprint data acquisition optimization model has the problem of low optimization performance parameters. Therefore, a data acquisition control optimization model based on neural network is proposed. This paper analyzes the advantages and disadvantages of neural network algorithm, combined with the sprint data collection optimization requirements, introduces BP neural network algorithm, based on this, uses multiple sensors, based on baud interval balance control to collect sprint data, applies BP neural network algorithm to compress, integrate and classify sprint data, realizes the sprint data collection and optimization. The experimental results show that the optimization performance parameters of the model are large, which fully shows that the model has good data acquisition optimization performance.

scholarly journals Emulating Balance Control Observed in Human Test Subjects with a Neural Network

2018 ◽  
pp. 200-212
Author(s):  
Wade W. Hilts ◽  
Nicholas S. Szczecinski ◽  
Roger D. Quinn ◽  
Alexander J. Hunt
Keyword(s):  
Neural Network ◽  
Balance Control

scholarly journals Adaptive Quadruped Balance Control for Dynamic Environments Using Maximum-Entropy Reinforcement Learning

Sensors ◽  
2021 ◽  
Vol 21 (17) ◽  
pp. 5907
Author(s):  
Haoran Sun ◽  
Tingting Fu ◽  
Yuanhuai Ling ◽  
Chaoming He
Keyword(s):  
Neural Network ◽  
Reinforcement Learning ◽  
Maximum Entropy ◽  
Maximum Entropy Method ◽  
Balance Control ◽  
External Disturbance ◽  
Control Policy ◽  
Dynamic Environments ◽  
Entropy Method ◽  
Analytical Models

External disturbance poses the primary threat to robot balance in dynamic environments. This paper provides a learning-based control architecture for quadrupedal self-balancing, which is adaptable to multiple unpredictable scenes of external continuous disturbance. Different from conventional methods which construct analytical models which explicitly reason the balancing process, our work utilized reinforcement learning and artificial neural network to avoid incomprehensible mathematical modeling. The control policy is composed of a neural network and a Tanh Gaussian policy, which implicitly establishes the fuzzy mapping from proprioceptive signals to action commands. During the training process, the maximum-entropy method (soft actor-critic algorithm) is employed to endow the policy with powerful exploration and generalization ability. The trained policy is validated in both simulations and realistic experiments with a customized quadruped robot. The results demonstrate that the policy can be easily transferred to the real world without elaborate configurations. Moreover, although this policy is trained in merely one specific vibration condition, it demonstrates robustness under conditions that were never encountered during training.

Sign in / Sign up

Export Citation Format

Share Document