Devlog 1: Project PathSense

BEHOLD the second project I am bringing to OutPost!

PathSense is a lightweight wearable vest designed for the visually impaired. It packs four Time-of-Flight (ToF) distance sensors, a GPS module, three vibration motors, a camera, and a Raspberry Pi 5 acting as the main command center.

Here is how the system handles the heavy lifting:

• The Hardware: The ToF sensors map out the environment up to four meters away, tracking obstructions to your left, right, and center.

• The AI (vague ykyk) : The camera runs a custom Python program for real-time object detection. The software uses the video feed as its primary source of truth. For example, if you are eating, the sensors might get blocked by your arms. The AI intelligently identifies your hands and ignores the sensor data during that time to avoid false alarms.

• The Haptic Feedback: If an obstacle appears on your left, the corresponding motor vibrates. The intensity changes depending on how close you are to the object.

• Audio Input: You can pair the vest with headphones for voice alerts. It will read out what is ahead, like letting you know a person is walking by or screaming “STOP!” if you approach a flight of stairs.Between the computer vision navigation and the GPS tracking, a person can confidently navigate both indoor spaces and busy city streets.

• Smart Navigation: As a massive upgrade, the Raspberry Pi 5 will run a custom navigation engine that combines active GPS coordinates with real-time computer vision pathfinding. Instead of just giving basic turn-by-turn directions, the system uses the camera feed to analyze the ground layout. It will actively look out for obstacles, steer the user away from rough or rocky terrain, and track the edges of sidewalks to ensure they never accidentally drift off the path.

The entire experience is designed to be fully voice-activated through a pair of headphones. The ultimate goal is to build it out like an interactive assistant (with possibility of combining it with the new and improved SIRI for fetching messages) , allowing the user to simply speak their destination and receive continuous, real-time auditory guidance as they walk.

As for today’s progress, there was only one main objective on my radar. I spent the hour designing a custom hardware HAT for the Raspberry Pi 5. This lets me drop standard JST connectors onto the board so all the sensors and motors can plug in cleanly without messy loose wires.

That is all for now. I will update everyone as soon as the grant clears and the manufacturing parts arrive!