Full Body Fitness XR

Project Overview

Working out in VR is incredible, but traditional VR fitness lacks personal touches and leaves certain muscle groups undertrained. Full Body Fitness XR solves this using Pico's full body trackers and SDK to create a personal training session that engages the entire body — including legs, core, and stabilizer muscles that most VR fitness apps miss.

Created for the Pico Redefining XR Hackathon, this project won Best Use of MR and secured 4th prize overall.

Technical Implementation

• Pico Full Body Tracking SDK — Real-time tracking of ankles, hips, and torso in addition to standard head/hand tracking, enabling true lower-body exercise
• Exercise Detection System — Custom algorithms that recognize and count repetitions of squats, lunges, kicks, and core twists based on tracker position and velocity curves
• MR Environment Blending — Real-world passthrough with virtual exercise guides overlaid, allowing users to see their actual environment while following virtual trainers
• Form Correction System — Real-time feedback when exercises are performed incorrectly, using tracker data to analyze joint angles and movement patterns
• Progressive Overload — Adaptive difficulty that increases intensity based on the user's performance and heart rate (estimated from movement data)
• Performance Optimization — Maintained 72 FPS on Pico 4 with full body tracking, a significant technical challenge given the additional tracking data pipeline

Design Approach

Most VR fitness games are upper-body only (boxing, beat sabers). Full Body Fitness XR was designed to bridge the gap between VR convenience and real workout effectiveness. The experience mimics a personal trainer session: warm-up, main workout with multiple exercise blocks, and cool-down. Each exercise is demonstrated by a virtual avatar, with the user's tracker skeleton overlaid for comparison.

The MR mode was critical — users need to see their surroundings to safely perform lunges and squats. The passthrough implementation keeps virtual elements anchored to the real world, with the trainer avatar appearing on a virtual pedestal that doesn't occlude floor space.

Key Results & Impact

• Winner — Best Use of MR at Pico Redefining XR Hackathon
• 4th Prize — Overall hackathon placement
• First-of-its-kind — True full-body workout in VR using consumer hardware
• 72 FPS — Stable performance with full body tracking on Pico 4

Tech Stack