Walking Direction Research Articles

Real-walk modelling: deep learning model for user mobility in virtual reality

This paper presents a study on modelling user free walk mobility in virtual reality (VR) art exhibition. The main objective is to investigate and model users’ mobility sequences during interactions with artwork in VR. We employ a range of machine learning (ML) techniques to define scenes of interest in VR, capturing user mobility patterns. Our approach utilises a long short-term memory (LSTM) model to effectively model and predict users’ future movements in VR environments, particularly in scenarios where clear walking paths and directions are not provided to participants. The DL model demonstrates high accuracy in predicting user movements, enabling a better understanding of audience interactions with the artwork. It opens avenues for developing new VR applications, such as community-based navigation, virtual art guides, and enhanced virtual audience engagement. The results highlight the potential for improved user engagement and effective navigation within virtual environments.

Redirected Walking on Omnidirectional Treadmill.

Redirected walking (RDW) and omnidirectional treadmill (ODT) are two effective solutions to the natural locomotion interface in virtual reality. ODT fully compresses the physical space and can be used as the integration carrier of all kinds of devices. However, the user experience varies in different directions of ODT, and the premise of interaction between users and integrated devices is a good match between virtual and real objects. RDW technology uses visual cues to guide the user's location in physical space. Based on this principle, combining RDW technology with ODT to guide the user's walking direction through visual cues can effectively improve user experience on ODT and make full use of various devices integrated on ODT. This paper explores the novel prospects of combining RDW technology with ODT and formally puts forward the concept of O-RDW (ODT-based RDW). Two baseline algorithms, i.e., OS2MD (ODT-based steer to multi-direction), and OS2MT (ODT-based steer to multi-target), are proposed to combine the merits of both RDW and ODT. With the help of the simulation environment, this paper quantitatively analyzes the applicable scenarios of the two algorithms and the influence of several main factors on the performance. Based on the conclusions of the simulation experiments, the two O-RDW algorithms are successfully applied in the practical application case of multi-target haptic feedback. Combined with the user study, the practicability and effectiveness of O-RDW technology in practical use are further verified.

F-RDW: Redirected Walking With Forecasting Future Position.

In order to serve better VR experiences to users, existing predictive methods of Redirected Walking (RDW) exploit future information to reduce the number of reset occurrences. However, such methods often impose a precondition during deployment, either in the virtual environment's layout or the user's walking direction, which constrains its universal applications. To tackle this challenge, we propose a mechanism F-RDW that is twofold: (1) forecasts the future information of a user in the virtual space without any assumptions by using the conventional method, and (2) fuse this information while maneuvering existing RDW methods. The backbone of the first step is an LSTM-based model that ingests the user's spatial and eye-tracking data to predict the user's future position in the virtual space, and the following step feeds those predicted values into existing RDW methods (such as MPCRed, S2C, TAPF, and ARC) while respecting their internal mechanism in applicable ways. The results of our simulation test and user study demonstrate the significance of future information when using RDW in small physical spaces or complex environments. We prove that the proposed mechanism significantly reduces the number of resets and increases the traveled distance between resets, hence augmenting the redirection performance of all RDW methods explored in this work. Our project and dataset are available at https://github.com/YonseiCGnA-VR/F-RDW.

Navigation System for Personal Mobility Vehicles Following a Cluster of Pedestrians in a Corridor Using Median of Candidate Vectors Observer

In recent years, personal mobility vehicles have been required to operate autonomously in places with numerous pedestrians. A navigation method using a single human-following scheme is used to avoid collision with pedestrians. However, in many cases, a single human-following method cannot be successfully used for guidance. In crowded places, pedestrians do not always keep walking in the desired direction a user wants to go, and the vehicle must change the target pedestrian frequently. Instead of following a single pedestrian, we propose a method for the vehicle to follow a cluster of pedestrians for stable and robust following. First, the pedestrians around the vehicle are detected by multiple RGB-D cameras, and the pedestrians are tracked using YOLO and Deep Sort. Pedestrians are classified according to their walking direction, and the cluster of pedestrians walking toward the goal is selected and followed. However, the position of pedestrian is sometimes lost in occlusions and the accuracy of the walking direction depends on the distance and pose detected by the sensors. A notable problem is that the cluster of pedestrians is unstable in the cluster following; therefore, a median of candidate vectors (MCV) observer is used to remove outliers caused by observation errors. The proposed method is applied to a scenario involving pedestrians walking toward an elevator hall in a building, and its effectiveness is verified through experiments.

Human Walking Direction Detection Using Wireless Signals, Machine and Deep Learning Algorithms.

The use of wireless signals for device-free activity recognition and precise indoor positioning has gained significant popularity recently. By taking advantage of the characteristics of the received signals, it is possible to establish a mapping between these signals and human activities. Existing approaches for detecting human walking direction have encountered challenges in adapting to changes in the surrounding environment or different people. In this paper, we propose a new approach that uses the channel state information of received wireless signals, a Hampel filter to remove the outliers, a Discrete wavelet transform to remove the noise and extract the important features, and finally, machine and deep learning algorithms to identify the walking direction for different people and in different environments. Through experimentation, we demonstrate that our approach achieved accuracy rates of 92.9%, 95.1%, and 89% in detecting human walking directions for untrained data collected from the classroom, the meeting room, and both rooms, respectively. Our results highlight the effectiveness of our approach even for people of different genders, heights, and environments, which utilizes machine and deep learning algorithms for low-cost deployment and device-free detection of human activities in indoor environments.

The flexible and distributed pressure sensor with 64 units for on-line gait recognition analysis

With the rapid development of flexible electronic technology, human gait recognition device is also developing in the direction of flexibility, refinement, and intelligence. In this paper, a gait recognition system with the advantages of being non-invasive and flexible is proposed and prepared to collect complex pressure information of human plantar and identify different gaits accurately. A flexible and distributed pressure sensor with a structure of eight rows and eight columns is designed and fabricated by the screen printing process and micro-nano manufacturing technology. The insole shape can be placed directly into a sneaker for easy wearing, while 64 pressure sensing units allow for more accurate detection of complex plantar pressure information. In addition, based on the “voltage mirror” principle and the feature threshold analysis method, a distributed pressure acquisition circuit and an online gait recognition software are developed respectively to realize time-sharing decoupling of complex output signals from multiple pressure sensing units and online real-time display of different gaits. Finally, Different human gait recognition experiments have also been carried out. The results show that the test error of the pressure sensing unit is better than 8.28%. The gait recognition system can realize the reliable and accurate online real-time identification of slow walking, fast walking, and different leaning directions, which shows great potential and value in the fields of medical diagnosis, health monitoring, gait correction and artificial intelligence.

A Versatile Topology-Optimized Compliant Actuator for Soft Robotic Gripper and Walking Robot.

The remarkable interaction capabilities of soft robots within various environments have captured substantial attention from researchers. In recent years, bionics has provided a rich inspiration for the design of soft robots. Nevertheless, predicting the locomotion of soft actuators and determining material layouts solely based on intuition or experience remain a formidable challenge. Previous actuators predominantly targeted separate applications, leading to elevated costs and diminished interchangeability. The objective of this article is to extract the common requirements of diverse application domains and develop a versatile compliant actuator. A mathematical model of the compliant mechanism is proposed under the framework of topology optimization, resulting in an optimal distribution of both structure and material. Through comparison with empirical and semioptimal designs, the results show that the proposed versatile actuator has the advantages of both stiffness and flexibility. We propose an associative design strategy for soft grippers and walking robots. The soft gripper can perfectly complete adaptive grasping of objects with varying sizes, shapes, and masses. The successful in-water gripping experiment underscores the robust cross-medium operational capabilities of the soft gripper. Notably, our experimental results show that the walking robot can move quickly for 5 cycles in 8.25 s and can guarantee the control accuracy of continuous motion. Moreover, the robot swiftly switches walking directions within a mere 0.45 s. The optimization and design strategy presented in this article can furnish novel insights for shaping the next generation of soft robots.

A pelvis-oriented margin of stability is robust against deviations in walking direction

The Margin of Stability (MOS) is often assessed relative to the intended, linear path of walking progression. When an unanticipated or irregular change in direction occurs, such as during a sudden turn or during activities of daily living, distinguishing the lateral from anteroposterior MOS can be challenging. The purpose of this study was to assess an anatomically orientated method of calculating the MOS using the pelvic orientation to define lateral and anteroposterior directions. We hypothesized that when straight walking was disrupted with a curved path, the pelvis-oriented MOS measure would be less variable compared to the global-oriented MOS measure. We recruited 16 unimpaired participants to walk at preferred and fast walking speeds along a straight walking path, as well as a path with an exaggerated, curvilinear deviation. We determined the within-subject mean and standard deviation of the anterior MOS at mid-swing and the posterior and lateral MOS at ipsilateral foot strike. For straight walking and curved walking separately, repeated measures factorial ANOVAs assessed the effects of model (global or pelvis-oriented), limb (left or right), and speed (preferred or fast) on these MOS values. Based on reduced variability during curved walking, the pelvis-oriented MOS was more robust to walking deviations than the globally defined MOS. In straight walking, the pelvis-oriented MOS was characterized by less lateral and more anterior stability with differences exacerbated by faster walking. These results suggest a pelvis-oriented MOS has utility when the path of progression is unknown or unclear.

A Lévy Flight-Inspired Random Walk Algorithm for Continuous Fitness Landscape Analysis

Heuristic algorithms are effective methods for solving complex optimization problems. The optimal algorithm selection for a specific optimization problem is a challenging task. Fitness landscape analysis (FLA) is used to understand the optimization problem's characteristics and help select the optimal algorithm. A random walk algorithm is an essential technique for FLA in continuous search space. However, most currently proposed random walk algorithms suffer from unbalanced sampling points. This article proposes a Lévy flight-based random walk (LRW) algorithm to address this problem. The Lévy flight is used to generate the proposed random walk algorithm's variable step size and direction. Some tests show that the proposed LRW algorithm performs better in the uniformity of sampling points. Besides, the authors analyze the fitness landscape of the CEC2017 benchmark functions using the proposed LRW algorithm. The experimental results indicate that the proposed LRW algorithm can better obtain the structural features of the landscape and has better stability than several other RW algorithms.

Modeling of pedestrian turning behavior and prediction of pedestrian density distribution

As urban public spaces attract more pedestrians, it is essential to prevent excessive pedestrian aggregation. Use of a walking behavior model is crucial to predict the distribution of pedestrian density. In earlier studies, the modeling of walking behavior generally focused on straight walking or aimless strolling. In recent years, more complex models of walking behavior, including turning behavior, have increasingly gained attention. Because the majority of models are calibrated using data gathered from experimental settings, research on the prediction of turning behavior of multi-scale walking passages in life scenarios remains in its development.The aim of this paper is to develop a turning behavior model suitable for multi-scale pedestrian passages. The expected walking direction was predicted by combining the global path planning model with a gradually changing expected direction with the local path planning model based on the attention field. Then, it was integrated with social force model to predict the pedestrian walking behavior and density distribution. Settings in the turning behavior model such as adaptively changing the expected walking direction according to the particular circumstance, approaching the inner curve upstream of the passage intersecting space, and appropriate pedestrian heterogeneity traits were found to help make predictions more accurately. The proposed turning behavior model is applicable to physical environments with different widths without recalibrating parameters and is suitable for predicting the distribution of pedestrian density in urban life scenarios with a variety of passage widths, which helps in the effective design and renovation of pedestrian space.

Watch your step: A pilot study of smartphone use effect on young females’ gait performance while walking up and down stairs and escalators

The increasingly ubiquitous use of smartphones has made distracted walking common, not only on flat ground, but also on stairs. Available information regarding changes in gait performance while walking and using a smartphone in different environments is still lacking. We aimed to investigate the differences in gait behavior and subjective walking confidence while walking up and down stairs and escalators, with and without smartphone use. A field experiment involving 32 female adults was conducted at a subway station. Gait parameters collected included step frequency, acceleration root mean square, step variability, step regularity, and step symmetry. The results showed that walking task, walking environment, and walking direction significantly affected gait performance and walking confidence. Overall, playing games or texting while walking down escalators resulted in the lowest walking confidence and the largest gait performance decrement: slower step frequency; reduced root mean square; decreased step regularity and step symmetry; and increased step variability. Step frequency, step variability, and step regularity significantly correlated with walking confidence. Smartphone use while walking on stairs and escalators significantly affects gait behavior and might increase the risk of falls. Interventions and prevention are needed to increase safety education and hazard warnings for the general population.

Life motion signals bias the perception of apparent motion direction.

Walking direction conveyed by biological motion (BM) cues, which humans are highly sensitive to since birth, can elicit involuntary shifts of attention to enhance the detection of static targets. Here, we demonstrated that such intrinsic sensitivity to walking direction could also modulate the direction perception of simultaneously presented dynamic stimuli. We showed that the perceived direction of apparent motion was biased towards the walking direction even though observers had been informed in advance that the walking direction of BM did not predict the apparent motion direction. In particular, rightward BM cues had an advantage over leftward BM cues in altering the perception of motion direction. Intriguingly, this perceptual bias disappeared when BM cues were shown inverted, or when the critical biological characteristics were removed from the cues. Critically, both the perceptual direction bias and the rightward advantage persisted even when only local BM cues were presented without any global configuration. Furthermore, the rightward advantage was found to be specific to social cues (i.e., BM), as it vanished when non-social cues (i.e., arrows) were utilized. Taken together, these findings support the existence of a specific processing mechanism for life motion signals and shed new light on their influences in a dynamic environment.

Life motion signals modulate visual working memory.

Previous research has demonstrated that biological motion (BM) cues can induce a reflexive attentional orienting effect, a phenomenon referred to as social attention. However, it remains undetermined whether BM cues can further affect higher-order cognitive processes, such as visual working memory (WM). By combining a modified central pre-cueing paradigm with a traditional WM change detection task, the current study investigated whether the walking direction of BM, as a non-predictive central cue, could modulate the encoding process of WM. Results revealed a significant improvement in WM performance for the items appearing at the location cued by the walking direction of BM. The observed effect disappeared when the BM cues were shown inverted, or when the critical biological characteristics of the cues were removed. Crucially, this effect could be extended to upright feet motion cues without global configuration, reflecting the key role of local BM signals in modulating WM. More importantly, such a BM-induced modulation effect was not observed with inanimate motion cues, although these cues can also elicit attentional effects. Our findings suggest that the attentional effect induced by life motion signals can penetrate to higher-order cognitive processes, and provide compelling evidence for the existence of "life motion detector" in the human brain from a high-level cognitive function perspective.

LUNCRW: Prediction of potential lncRNA-disease associations based on unbalanced neighborhood constraint random walk

Accumulating evidence suggests that long non-coding RNAs (lncRNAs) are associated with various complex human diseases. They can serve as disease biomarkers and hold considerable promise for the prevention and treatment of various diseases. The traditional random walk algorithms generally exclude the effect of non-neighboring nodes on random walking. In order to overcome the issue, the neighborhood constraint (NC) approach is proposed in this study for regulating the direction of the random walk by computing the effects of both neighboring nodes and non-neighboring nodes. Then the association matrix is updated by matrix multiplication for minimizing the effect of the false negative data. The heterogeneous lncRNA-disease network is finally analyzed using an unbalanced random walk method for predicting the potential lncRNA-disease associations. The LUNCRW model is therefore developed for predicting potential lncRNA-disease associations. The area under the curve (AUC) values of the LUNCRW model in leave-one-out cross-validation and five-fold cross-validation were 0.951 and 0.9486 ± 0.0011, respectively. Data from published case studies on three diseases, including squamous cell carcinoma, hepatocellular carcinoma, and renal cell carcinoma, confirmed the predictive potential of the LUNCRW model. Altogether, the findings indicated that the performance of the LUNCRW method is superior to that of existing methods in predicting potential lncRNA-disease associations.

The FreeD module's lateral translation timing in the gait robot Lokomat: a manual adaptation is necessary.

Pelvic and trunk movements are often restricted in stationary robotic gait trainers. The optional FreeD module of the driven gait orthosis Lokomat offers a combined, guided lateral translation and transverse rotation of the pelvis and may therefore support weight shifting during walking. However, from clinical experience, it seems that the default setting of this timing does not correspond well with the timing of the physiological pelvic movement during the gait cycle. In the software, a manual adaptation of the lateral translation's timing with respect to the gait cycle is possible. The aim of this study was to investigate if such an offset is indeed present and if a manual adaptation by the therapist can improve the timing towards a more physiological pattern comparable to physiological overground walking. Children and adolescents with neurologic gait disorders and a Gross Motor Function Classification System level I-IV completed two different walking conditions (FreeD Default and FreeD Time Offset) in the Lokomat. The medio-lateral center of mass positions were calculated from RGB-Depth video recordings with a marker-less motion capture algorithm. Data of 22 patients (mean age: 12 ± 3 years) were analyzed. Kinematic analyses showed that in the FreeD Default condition, the maximum lateral center of mass excursion occurred too early. In the FreeD Time Offset condition, the manual adaptation by the therapists led to a delay of the maximum center of mass displacement by 8.2% in the first phase of the gait cycle and by 4.9% in the second phase of the gait cycle compared to the FreeD Default condition. The maximum lateral center of mass excursion was closer to that during physiological overground walking in the FreeD Time Offset condition than in the FreeD Default condition. A manual adaptation of the timing of the FreeD module in the Lokomat shifts pelvis kinematics in a direction of physiological overground walking. We recommend therapists to use this FreeD Time Offset function to adjust the phase of weight shifting for each patient individually to optimize the kinematic walking pattern when a restorative therapy approach is adopted.

Adaptation to walking direction in biological motion.

The direction that we see another person walking provides us with an important cue to their intentions, but little is known about how the brain encodes walking direction across a neuronal population. The current study used an adaptation technique to investigate the sensory coding of perceived walking direction. We measured perceived walking direction of point-light stimuli before and after adaptation, and found that adaptation to a specific walking direction resulted in repulsive perceptual aftereffects. The magnitude of these aftereffects was tuned to the walking direction of the adaptor relative to the test, with local repulsion of perceived walking direction for test stimuli oriented on either side of the adapted walking direction. The specific tuning profiles that we observed are well explained by a population-coding model, in which perceived walking direction is coded in terms of the relative activity across a bank of sensory channels with peak tuning distributed across the full 360° range of walking directions. Further experiments showed specificity in how horizontal (azimuth) walking direction is coded when moving away from the observer compared to when moving toward the observer. Moreover, there was clear specificity in these perceptual aftereffects for walking direction compared to a nonbiological form of 3D motion (a rotating sphere). These results indicate the existence of neural mechanisms in the human visual system tuned to specific walking directions, provide insight into the number of sensory channels and how their responses are combined to encode walking direction, and demonstrate the specificity of adaptation to biological motion. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Logical Topology Inference via CPGCN Joint Optimizing With Pedestrian Re-Id.

With the rise of artificial intelligence, deep learning has become the main research method of pedestrian recognition re-identification (re-id). However, most of the existing researches usually just determine the retrieval order based on the geographical location of cameras, which ignore the spatio-temporal logic characteristics of pedestrian flow. Furthermore, most of these methods rely on common object detection to detect and match pedestrians directly, which will separate the logical connection between videos from different cameras. In this research, a novel pedestrian re-identification model assisted by logical topological inference is proposed, which includes: 1) a joint optimization mechanism of pedestrian re-identification and multicamera logical topology inference, which makes the multicamera logical topology provides the retrieval order and the confidence for re-identification. And meanwhile, the results of pedestrian re-identification as a feedback modify logical topological inference; 2) a dynamic spatio-temporal information driving logical topology inference method via conditional probability graph convolution network (CPGCN) with random forest-based transition activation mechanism (RF-TAM) is proposed, which focuses on the pedestrian's walking direction at different moments; and 3) a pedestrian group cluster graph convolution network (GC-GCN) is designed to measure the correlation between embedded pedestrian features. Some experimental analyses and real scene experiments on datasets CUHK-SYSU, PRW, SLP, and UJS-reID indicate that the designed model can achieve a better logical topology inference with an accuracy of 87.3% and achieve the top-1 accuracy of 77.4% and the mAP accuracy of 74.3% for pedestrian re-identification.

Happy is stronger than sad: Emotional information modulates social attention.

Social directional cues (e.g., gaze direction; walking direction) can trigger reflexive attentional orienting, a phenomenon known as social attention. Here, we examined whether this reflexive social attention could be modulated by the emotional content embedded in social cues. By introducing emotional (happy and sad) biological motion (BM) stimuli to the modified central cuing paradigm, we found that the happy but not the sad emotional gait could significantly boost attentional orienting effect relative to the neutral gait. Critically, this "happiness advantage" effect could be extended to social attention induced by gaze. Furthermore, the observed differential emotional modulations could not be simply explained by low-level physical differences between the emotional stimuli, as inverted social cues (i.e., BM and face) failed to produce such modulation effects. Overall, these findings highlight the role of emotional information in modulating the processing of social signals, and further suggest the existence of a general emotional modulation on social attention triggered by different types of social signals. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

A Privacy-Protecting Step-Level Walking Direction Detection Algorithm Based on Floor Vibration

We present an algorithm and measurement system to detect the walking direction of persons based on ground vibrations. The approach is privacy-preserving because it solely relies on piezoelectric sensors built into the floor. Therefore, our system can be used in areas where cameras are not allowed or cannot capture the entire area. We present and compare our two innovative methods to analyze the ground vibrations caused by footsteps: the multi-peaks average algorithm (MPAA) and the multi-peaks averaged feature with a deep neural network-based classifier (MPAF-DNNC). MPAA judges the walking direction of pedestrians by analyzing the time-space relationship of at least two consecutive footstep vibration signals from multiple sensors. MPAF-DNNC receives multi-peaks averaged feature as input and uses a deep neural network-based classifier to judge walking direction. Our experiments and evaluation show that our system can correctly determine the walking direction based on only 3 input step events and provides an average F1 score of 0.97. When more than 5 step events are inputted, the proposed system can correctly determine the walking direction with an average F1 score of 1.00.

Robust and reusable self-organized locomotion of legged robots under adaptive physical and neural communications.

Animals such as cattle can achieve versatile and elegant behaviors through automatic sensorimotor coordination. Their self-organized movements convey an impression of adaptability, robustness, and motor memory. However, the adaptive mechanisms underlying such natural abilities of these animals have not been completely realized in artificial legged systems. Hence, we propose adaptive neural control that can mimic these abilities through adaptive physical and neural communications. The control algorithm consists of distributed local central pattern generator (CPG)-based neural circuits for generating basic leg movements, an adaptive sensory feedback mechanism for generating self-organized phase relationships among the local CPG circuits, and an adaptive neural coupling mechanism for transferring and storing the formed phase relationships (a gait pattern) into the neural structure. The adaptive neural control was evaluated in experiments using a quadruped robot. The adaptive neural control enabled the robot to 1) rapidly and automatically form its gait (i.e., self-organized locomotion) within a few seconds, 2) memorize the gait for later recovery, and 3) robustly walk, even when a sensory feedback malfunction occurs. It also enabled maneuverability, with the robot being able to change its walking speed and direction. Moreover, implementing adaptive physical and neural communications provided an opportunity for understanding the mechanism of motor memory formation. Overall, this study demonstrates that the integration of the two forms of communications through adaptive neural control is a powerful way to achieve robust and reusable self-organized locomotion in legged robots.