Gesture Recognition with Hidden Markov Models to Enable Multi-modal Haptic Feedback

Presentation of more haptic information to improve the users interactive capability is an important and challenging task in the virtual environment. As an emerging haptic feedback technology, electrovibration is an underlying solution to enhance the systems interactivity and improve user experience. However, such a solution is rarely studied in a virtual environment. In this work, we explore a new VR interaction method based on electrovibration technology with a touch screen and conduct evaluations about it. The key idea is to incorporate a set of manipulation gestures and three types of electrovibration in the VR interaction to help users acquire different kinds of tactile perception in the virtual manipulation. We present the evaluation in which we compare user performance measured first in a Fitts law task to evaluate different electrovibration types and then in a virtual office application to assess the interactive user interface. Our results show that the precision of interactions is significantly improved with the electrovibration haptic feedback. To the best of our knowledge, this is the first work to introduce electrovibration haptic feedback into the VR human-computer interaction and our work enlightens the potential of the electrovibration touchscreen-based interaction in virtual environments.

In this demonstration, we introduce Face/On, an embedded feedback device that leverages the contact area between the user's face and a virtual reality (VR) head-mounted display (HMD) to provide rich haptic feedback in virtual environments (VEs). Head-worn haptic feedback devices have been explored in previous work to provide directional cues via grids of actuators and localized feedback on the users' skin. Most of these solutions were immersion breaking due to their encumbering and uncomfortable design and build around a single actuator type, thus limiting the overall fidelity and flexibility of the haptic feedback. We present Face/On, a VR HMD face cushion with three types of discreetly embedded actuators that provide rich haptic feedback without encumbering users with invasive instrumentation on the body. By combining vibro-tactile and thermal feedback with electrical muscle stimulation (EMS), Face/On can simulate a wide range of scenarios and benefit from synergy effects between these feedback types.

Sound is an important, but often neglected, component for creating a self-motion illusion (vection) in Virtual Reality applications, for example, motion simulators. Apart from auditory motion cues, sound can provide contextual information representing self-motion in a virtual environment. In two experiments we investigated the benefits of hearing an engine sound when presenting auditory (Experiment 1) or auditory-vibrotactile (Experiment 2) virtual environments inducing linear vection. The addition of the engine sound to the auditory scene significantly enhanced subjective ratings of vection intensity in Experiment 1 and vection onset times but not subjective ratings in Experiment 2. Further analysis using individual imagery vividness scores showed that this disparity between vection measures was created by participants with higher kinesthetic imagery. On the other hand, for participants with lower kinesthetic imagery scores, the engine sound enhanced vection sensation in both experiments. A high correlation with participants' kinesthetic imagery vividness scores suggests the influence of a first person perspective in the perception of the engine sound. We hypothesize that self-motion sounds (e.g., the sound of footsteps, engine sound) represent a specific type of acoustic body-centered feedback in virtual environments. Therefore, the results may contribute to a better understanding of the role of self-representation sounds (sonic self-avatars), in virtual and augmented environments.

The usability of virtual reality (VR) controller interfaces are often complex and difficult for first time users. Most controllers provide minimal feedback which relegates the potential for heightened interaction and feedback within virtual experiences. This research explores how haptic technology systems partnered with VR can deliver immersive interactions between user and virtual environment (VE). This research involves the development of a haptic glove interface prototype that incorporates a force feedback and vibrotactile feedback system. It focuses on determining a workflow that communicates in real-time user interaction and environmental feedback using Unreal Engine and the produced haptic glove system. Testing and calibrating the prototype feedback system provided a baseline for developers to rationalise and improve accuracy of current real-time virtual feedback systems. The evaluation of this research in industry unfolds new technical knowledge for implementing a wider range of haptic technologies within VR. This further development would involve reviewing the usability and interaction standards for VR users in the design process.

Human gesture recognition is an area, which has been studied thoroughly in recent years,and close to100% recognition rates in restricted environments have been achieved, often either with single separated gestures in the input stream, or with computationally intensive systems. The results are unfortunately not as striking, when it comes to a continuous stream of gestures. In this paper we introduce a hierarchical system for gesture recognition for use in a gaming setting, with a continuous stream of data. Layer 1 is based on Nearest Neighbor Search and layer 2 uses Hidden Markov Models. The system uses features that are computed from Microsoft Kinect skeletons. We propose a new set of features, the relative angles of the limbs from Kinect's axes to use in NNS. The new features show a 10 percent point increase in precision when compared with features from previously published results. We also propose a way of attributing recognised gestures with a force attribute, for use in gaming. The recognition rate in layer 1 is 68.2%, with an even higher rate for simple gestures. Layer 2 reduces the noise and has aaverage recognition rate of 85.1%. When some simple constraints are added we reach a precision of 90.5% with a recall of 91.4%.

In this article, a real-time dynamic finger gesture recognition using a soft sensor embedded data glove is presented, which measures the metacarpophalangeal (MCP) and proximal interphalangeal (PIP) joint angles of five fingers. In the gesture recognition field, a challenging problem is that of separating meaningful dynamic gestures from a continuous data stream. Unconscious hand motions or sudden tremors, which can easily lead to segmentation ambiguity, makes this problem difficult. Furthermore, the hand shapes and speeds of users differ when performing the same dynamic gesture, and even those made by one user often vary. To solve the problem of separating meaningful dynamic gestures, we propose a deep learning-based gesture spotting algorithm that detects the start/end of a gesture sequence in a continuous data stream. The gesture spotting algorithm takes window data and estimates a scalar value named gesture progress sequence (GPS). GPS is a quantity that represents gesture progress. Moreover, to solve the gesture variation problem, we propose a sequence simplification algorithm and a deep learning-based gesture recognition algorithm. The proposed three algorithms (gesture spotting algorithm, sequence simplification algorithm, and gesture recognition algorithm) are unified into the real-time gesture recognition system and the system was tested with 11 dynamic finger gestures in real-time. The proposed system took only 6 ms to estimate a GPS and no more than 12 ms to recognize the completed gesture in real-time.

Vibrotactile feedback in virtual motor learning: A systematic review

Current immersive Virtual Reality (VR) system strategies do not fully support dynamic Human Computer Interaction (HCI) and since there is a growing need for better immersion, due consideration should be given to integrate additional modalities for improved HCI. While feedback in Virtual Environments (VE) is predominantly provided to the user through the visual and auditory channels, additional modalities such as haptics can increase the sense of presence and efficiency in VE simulations. Haptic interfaces can enhance the VE interaction by enabling users to “touch” and “feel” virtual objects that are simulated in the environment. This paper examines the reasons behind its integration based on the limitations of present immersive projection system.

Recognizing dynamic gestures from a continuous data stream has been treated as a difficult task. One challenge is the gesture spotting problem, or segmentation of a gesture pattern from a data stream containing consecutive gestures. Although various types of gesture spotting strategies have been introduced so far, all methods has its own limitations to be applied to real world situations such as sign language conversation. One of the conditions that must be satisfied for practical applications is to detect repetitive gestures for emphasizing a certain meaning.In the view of gesture spotting, repetitive gestures are hard to be detected because similar starting and ending moments are repeated without constraints of the number of repetitions. We introduce a data glove-based real-time dynamic gesture recognition system that covers both repetitive and nonrepetitive gestures. The proposed system consists of three steps: Gesture spotting, gesture sequence compression, and gesture recognition. For gesture spotting, we define a quantity termed as a gesture progress sequence, which quantifies a progress of a gesture using numbers from 0 to 1. Gesture sequence compression removes variations including the number of repetitions that seek to impart the same message. At the gesture recognition step, the compressed gesture patterns are classified. The proposed system was evaluated using 17 American sign language (ASL) gestures and some ASL sentences.

As worldwide population is aging and concentrating in cities, designing a pedestrian navigation aid adapted to elderly specificities and needs could help in preserving their autonomy, as their cognitive and perceptive abilities may decline with age. This study was aimed at comparing visual, auditory and haptic guidance feedbacks among older pedestrians, taking into account their navigation performance and user experience. A virtual environment was used to measure time to destination and rate of correct responses. Post-activity interviews were used to access participants' perceptions, feelings and hesitations in a qualitative way. Results showed great performance and user experience for visual guidance, whereas performance and user experience were poor for haptic guidance.

This paper presents a human-scale virtual environment (VE) with haptic feedback along with two experiments performed in the context of product design. The user interacts with a virtual mock-up using a large-scale bimanual string-based haptic interface called SPIDAR (Space Interface Device for Artificial Reality). An original self-calibration method is proposed. A vibro-tactile glove was developed and integrated to the SPIDAR to provide tactile cues to the operator. The purpose of the first experiment was: (1) to examine the effect of tactile feedback in a task involving reach-and-touch of different parts of a digital mock-up, and (2) to investigate the use of sensory substitution in such tasks. The second experiment aimed to investigate the effect of visual and auditory feedback in a car-light maintenance task. Results of the first experiment indicate that the users could easily and quickly access and finely touch the different parts of the digital mock-up when sensory feedback (either visual, auditory, or tactile) was present. Results of the of the second experiment show that visual and auditory feedbacks improve average placement accuracy by about 54 % and 60% respectively compared to the open loop case

A low-power and real-time hidden Markov model (HMM) accelerator is proposed for gesture user interfaces on wearable smart devices. HMM algorithm is widely used for sequence recognitions such as speech recognition and gesture recognition due to its best-in-class recognition accuracy. However, the HMM algorithm incorporates high computational complexity and requires massive memory bandwidth for sequence matches. There have been studies on hardware acceleration of the HMM algorithm to resolve these issues, but they were focused on the speech recognition and did not incorporate the motion orientation capability required for the gesture recognition case. In this paper, we propose an HMM accelerator incorporating the motion orientation block for gesture recognitions on wearable devices. Binary search is exploited in the motion orientation to avoid the division and arctangent associated with the orientation and reduce its arithmetic complexity. In addition, gesture models are clustered in the gesture database to save the memory bandwidth by reducing memory transactions. Moreover, logarithmic arithmetic is used in Viterbi decoding in the HMM for more reduction in its complexity. Thanks to these schemes, this work achieves 25.6% power reduction compared with a plain hardware implementation of the gesture recognizing HMM.

Internet of Things (IoT) systems include many smart devices that continuously generate massive spatio-temporal data, which can be difficult to process. These continuous data streams need to be stored smartly so that query searches are efficient. In this work, we propose an efficient method, in the fog-cloud computing architecture, to index continuous and heterogeneous data streams in metric space. This method divides the fog layer into three levels: clustering, clusters processing and indexing. The Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm is used to group the data from each stream into homogeneous clusters at the clustering fog level. Each cluster in the first data stream is stored in the clusters processing fog level and indexed directly in the indexing fog level in a Binary tree with Hyperplane (BH tree). The indexing of clusters in the subsequent data stream is determined by the coefficient of variation (CV) value of the union of the new cluster with the existing clusters in the cluster processing fog layer. An analysis and comparison of our experimental results with other results in the literature demonstrated the effectiveness of the CV method in reducing energy consumption during BH tree construction, as well as reducing the search time and energy consumption during a k Nearest Neighbor (kNN) parallel query search.

Mastering ballroom dance requires learning motor skills such as synchronizing the body motion to a tempo as well as memorizing long sequences of dance steps. Traditionally, the acquisition of these skills is guided by an instructor, who adapts the training to the level of the trainee. However, due to the measures that are taken globally against the COVID19, a great number of training facilities have partially or completely restricted access to their trainers. To facilitate the practice of ballroom dance without the need of a human instructor nor human contact, this paper presents a mixed-reality human-machine interface that provides part practice of ballroom dance. It achieves it by fractionalizing the dance into only the footwork and conveying the tempo, position trajectory, and velocity trajectory with multimodal feedback. The multimodal feedback is generated with a floor projection, vibrotactile actuators, and a metronome, which provide visual, haptic, and auditory feedback. In an experiment with 15 participants, results suggest that the motor learning of dance skills occurs when visual and haptic feedback is present, whereas it is further improved when the three modalities (visual, haptic, and auditory) are employed during training.

In order to provide relevant information to mobile users, such as workers engaging in the manual tasks of maintenance and assembly, a wearable computer requires information about the user's specific activities. This work focuses on the recognition of activities that are characterized by a hand motion and an accompanying sound. Suitable activities can be found in assembly and maintenance work. Here, we provide an initial exploration into the problem domain of continuous activity recognition using on-body sensing. We use a mock "wood workshop" assembly task to ground our investigation. We describe a method for the continuous recognition of activities (sawing, hammering, filing, drilling, grinding, sanding, opening a drawer, tightening a vise, and turning a screwdriver) using microphones and three-axis accelerometers mounted at two positions on the user's arms. Potentially "interesting" activities are segmented from continuous streams of data using an analysis of the sound intensity detected at the two different locations. Activity classification is then performed on these detected segments using linear discriminant analysis (LDA) on the sound channel and hidden Markov models (HMMs) on the acceleration data. Four different methods at classifier fusion are compared for improving these classifications. Using user-dependent training, we obtain continuous average recall and precision rates (for positive activities) of 78 percent and 74 percent, respectively. Using user-independent training (leave-one-out across five users), we obtain recall rates of 66 percent and precision rates of 63 percent. In isolation, these activities were recognized with accuracies of 98 percent, 87 percent, and 95 percent for the user-dependent, user-independent, and user-adapted cases, respectively.