The high number of fatal crashes caused by driver drowsiness highlights the need for developing reliable drowsiness detection methods. An ideal driver drowsiness detection system should estimate multiple levels of drowsiness accurately without intervening in the driving task. This paper proposes a multi-level drowsiness detection system by a deep neural network-based classification system using a combination of electrocardiogram and respiration signals. The proposed method is based on a combination of convolutional neural networks (CNNs) and long short-term memory (LSTM) networks for classifying drowsiness by concurrently using heart rate variability (HRV), power spectral density of HRV, and respiration rate signal as inputs. Two models, a CNN-based model and a hybrid CNN-LSTM-based model were used for multi-level classifications. The performance of the proposed method was evaluated on experimental data collected from 30 subjects in a simulated driving environment. The performance and the results of both models are presented and compared. The best performance for both three-level and five-level drowsiness classifications was achieved by the CNN-LSTM model. The results indicate that the three-level and five-level classifications of drowsiness can be achieved with 91 and 67% accuracy, respectively.
Detecting human movement is crucial for the control of lower limb wearable robotics designed to assist daily activities or rehabilitation tasks. Sensorized insoles present a viable option for extracting control inputs, such as gait events and the corresponding phases, essential for regulating the magnitude and timing of assistance. Given their highly sensitive piezoelectric response to dynamic loading, ferroelectrets emerge as a cost-effective solution for customizing sensors suitable for these autonomous systems. Within this study, an insole with four ferroelectret sensors is 3D-printed monolithically from polylactic acid (PLA) onto bulk films of the same material through seamless thermal fusion. Sensor and insole are characterized through a testing machine and by conducting human walking experiments on an instrumented treadmill. The testing machine results indicate suitable sensor performance for the application in wearable robotics concerning the sensitivity, minimal detectable change, hysteresis, drift, and repeatability. Walking experiments reveal the insole’s capability to detect gait events such as heel strikes with minimal variability and on average 16 ms faster compared to the reference of vertical ground reaction forces across all walking speeds above 1 m/s. The peak sensor outputs strongly relate to the reference while both exhibit a linear (R-squared > 95%) increase corresponding to walking speed. In conclusion, study findings demonstrate the feasibility of PLA-based ferroelectrets as customized insole sensors for event detection in gait analysis, enabling assessment of human biomechanics with minimal impact on the natural gait and control of autonomous wearable robotics, such as exoskeletons.
Biomechanical models with different levels of complexity are of advantage to understand the underlying principles of legged locomotion. Following a minimalistic approach of gradually increasing model complexity based onTemplate & Anchorconcept, in this paper, a spring-loaded inverted pendulum-based walking model is extended by a rigid trunk, hip muscles and reflex control, called nmF (neuromuscular force modulated compliant hip) model. Our control strategy includes leg force feedback to activate hip muscles (originated from the FMCH approach), and a discrete linear quadratic regulator for adapting muscle reflexes. The nmF model demonstrates human-like walking kinematic and dynamic features such as the virtual pendulum (VP) concept, inherited from the FMCH model. Moreover, the robustness against postural perturbations is two times higher in the nmF model compared to the FMCH model and even further increased in the adaptive nmF model. This is due to the intrinsic muscle dynamics and the tuning of the reflex gains. With this, we demonstrate, for the first time, the evolution of mechanical template models (e.g. VP concept) to a more physiological level (nmF model). This shows that the template model can be successfully used to design and control robust locomotor systems with more realistic system behaviours.
11 May 2022 in Journal of Intelligent & Robotic Systems
In order to approach the performance of biological locomotion in legged robots, better integration between body design and control is required. In that respect, understanding the mechanics and control of human locomotion will help us build legged robots with comparable efficient performance. From another perspective, developing bioinspired robots can also improve our understanding of human locomotion. In this work, we create a bioinspired robot with a blended physical and virtual impedance control to configure the robot’s mechatronic setup. We consider human neural control and musculoskeletal system a blueprint for a hopping robot. The hybrid electric-pneumatic actuator (EPA) presents an artificial copy of this biological system to implement the blended control. By defining efficacy as a metric that encompasses both performance and efficiency, we demonstrate that incorporating a simple force-based control besides constant pressure pneumatic artificial muscles (PAM) alone can increase the efficiency up to 21% in simulations and 7% in experiments with the 2-segmented EPA-hopper robot. Also, we show that with proper adjustment of the force-based controller and the PAMs, efficacy can be further increased to 41%. Finally, experimental results with the 3-segmented EPA-hopper robot and comparisons with human hopping confirm the extendability of the proposed methods to more complex robots.
There is a long history demonstrating humans’ tendency to create artificial copies of living creatures. For moving machines called robots, actuators play a key role in developing human-like movements. Among different types of actuation, PAMs (pneumatic artificial muscles) are known as the most similar ones to biological muscles. In addition to similarities in force generation mechanism (tension based), the well-accepted argumentation from Klute et al., states that the PAM force-length (f l ) behavior is close to biological muscles, while the force-velocity (f v ) pattern is different. Using the multiplicative formulation of the pressure (as an activation term), f l and f v beside an additive passive parallel elastic element, we present a new model of PAM. This muscle-based model can predict PAM dynamic behaviors with high precision. With a second experiment on a two-segmented leg, the proposed model is verified to predict the generated forces of PAMs in an antagonistic arrangement. Such a dynamic muscle-like model of artificial muscles can be used for the design and control of legged robots to generate robust, efficient and versatile gaits.
1 September 2021 in Robotics and Autonomous Systems