Experimental Evaluations

  1. Measured UWB inaccuracy was reported as 20-30 cm (10 cm in datasheet)
  2. Inaccuracy due to the UWB anchors placement made navigation hard.
  3. Pose estimation factor of ~20 cm in x and y coordinates was calculated and added to properly find the user’s body.
  4. Based on testing the wearable on 3 participants, navigation commands were most accurate once the user’s arm was by his side with the BNO055 pointing downward.
  5. Bluetooth Beacon successfully addressed the inaccuracy due to UWB’s error except for the cases where it was placed next to Microwave.

Strengths

  • The shortest path calculation gives an advantage over other products
  • Utilizing Bluetooth beacon to reduce the error rate due to UWB modules helps to solve the problem of the last centimeters of precision
  • User friendly App features such as “adding new locations” is useful to extend the functionality of the system developed.
  • Properly calculating the Pose and taking into account the user’s body median plane gives the system more precision.

Limitation and Future Works

  • DirectMe lacks obstacle avoidance technology.
  • Better configuration of UWB anchors would help with positioning accuracy.
  • DirectMe requires more implicit commands so it's easier for ordinary users to work with.
  • Once obstacle avoidance added, the search algorithm would also needs to be optimized.
  • As the Bluetooth beacon failed in some scenarios, DirectMe needs a better way of getting precise location data in order to be more robust.

Future Works

One good way in terms of future works would be to integrate our with current design with smart canes. Embedding DirectMe on WeWalk smart cane for example, it will stisfy our obstacle avoidance need while adding indoor navigation to one of the best innovations of 2019 according to Time Magazine.

Related Works:


  • Minuet: Multimodal Interaction with an Internet of Things

    • This was our main inspiration. As we almost followed the exact same setup of they did in their paper for the purpose of indoor localization.

  • Indoor Positioning for Visually Impaired People Based on Smartphones

    • This paper mainly discussed how the current sensors in smartphone can be utilized for the purpose of indoor navigation. We used this one mainly for expanding our ideas and referred to it in future works and the limitations of our current model.

  • Analysis of a UWB Indoor Positioning System Based on Received Signal Strength

    • This way our mainly idea of adding signal strength analysis to enhance the error rate UWB anchors. However due to lack of time we used bluetooth Beacon as explained in System Overview. However this paper remains an interesting idea to focus on for later projects.

  • Pozyx: Localization via UWB

    • This one was a commercial product with very precise claimed and report accuracy. The provided software in their products provide a heat map of how well the UWB system perfom given a specific set of placement of anchors.