Abstract:Redirected walking (RDW) allows users to explore vast virtual spaces by walking in confined real spaces, yet suffers from frequent boundary collisions due to physical constraints. The major solution is to use the reset strategy to steer users away from boundaries. However, most reset methods guide users to fixed spots or follow constant patterns, neglecting spatial features and users' movement trends. In this paper, we propose an innovative predictive reset method based on spatial probability density distribution to jointly involve impacts of spatial feature and walking intention for forecasting the user's possible positional distribution, and thereby determines the optimal reset direction by maximizing walking expectation. Given a space, we calculate the stationary layout energy to indicate traveling difficulties of all positions. Meanwhile, we exploit a novel intention inference model to anticipate the probability distribution of the user's presence across adjacent positions. Furthermore, we incorporate the obstacle energy attenuation to predict the obstacle avoidance behaviors. All aforementioned factors are amalgamated into a potential region energy map, and then we integrate energy maps of virtual and real spaces into a fusion energy map to enable the prediction considering both spaces simultaneously. Thus, the optimal reset direction is derived by maximizing the fusion energy. Simulation and user studies are conducted on a broad dataset containing plentiful virtual and real spaces. The results demonstrate that our method effectively reduces the physical collisions and increase the continuous walking distance compared to prevalent reset methods, while exhibiting superior applicability when combined with various RDW controllers. The source code and dataset are available at https://github.com/huiyuroy/Return2MaxPotentialEnergy.

OpenRDW: A Redirected Walking Library and Benchmark with Multi-User, Learning-based Functionalities and State-of-the-art Algorithms

Redirected Walking Based on Historical User Walking Data.

FREE-RDW: A Multi-user Redirected Walking Method for Supporting Non-forward Steps

Redirected Walking for Exploring Immersive Virtual Spaces with HMD: A Comprehensive Review and Recent Advances

A Real-Time Randomized Navigation Method for Biped Robot

A mapping-based redirected walking algorithm for large-scale VR

F-RDW: Redirected Walking with Forecasting Future Position

VDBblox: Accurate and Efficient Distance Fields for Path Planning and Mesh Reconstruction

Adaptive Redirection: A Context-Aware Redirected Walking Meta-Strategy

Comparing Four Approaches to Generalized Redirected Walking: Simulation and Live User Data

Performance of Redirected Walking Algorithms in a Constrained Virtual World

SafeRDW: Keep VR Users Safe when Jumping Using Redirected Walking

Multi-User Redirected Walking in Separate Physical Spaces for Online VR Scenarios

Transferable Virtual-Physical Environmental Alignment with Redirected Walking

Enabling Predictive Redirection Reset Based on Virtual-Real Spatial Probability Density Distributions

Optimizing Constrained-Environment Redirected Walking Instructions Using Search Techniques

BiRD: Using Bidirectional Rotation Gain Differences to Redirect Users During Back-and-forth Head Turns in Walking

Optimal Pose Guided Redirected Walking with Pose Score Precomputation

Overcoming Spatial Constraints in VR: A Survey of Redirected Walking Techniques

Adaptive Optimization Algorithm for Resetting Techniques in Obstacle-Ridden Environments