Apple received 2 granted patents this quarter across 2 categories: VR & AR (1) and AI & Machine Learning (1).
The VR & AR patent covers frame rate extrapolation technology that uses motion sensors to generate intermediate frames, reducing latency and flickering in enhanced reality experiences. In AI & Machine Learning, Apple patented an agent-agnostic motion planner that generates movement paths for virtual characters across different XR/VR applications without requiring custom motion planners for each character.
Apple's single VR & AR patent tackles the challenge of smooth visual rendering in headsets by creating intermediate frames between actual rendered ones. The system uses motion sensor data to predict where objects should appear in those in-between moments, employing a clever dual approach that processes the user's central field of view differently from the periphery. At the center, it applies a simple single-depth transformation across the entire visible plane, while the edges receive more complex pixel-by-pixel depth calculations that account for varying distances. A blending function smoothly transitions between these two methods, eliminating the peripheral flicker that plagues simpler approaches without the performance hit of processing every pixel at maximum complexity.
The AI & Machine Learning patent addresses the tedious problem of building separate movement systems for every virtual character type in XR and VR environments. Instead of requiring developers to create custom motion planners for each agent, the system analyzes how a character's existing motion controller works at runtime and adapts its path planning accordingly. It generates verified movement paths using rapidly-exploring random trees combined with machine-learned motion models, all while respecting each agent's unique control scheme and physical capabilities.