top of page
PEV-final.jpg

PEV

A lightweight autonomous vehicle for urban areas

Emerging urban contexts and societal aspirations​

PEV stands for Persuasive Electric Vehicle and is a research project of the MIT Media Lab. It aims to solve urban mobility challenges with a healthy, convenient, sustainable alternative to cars. The PEV is a low-cost, agile, shared-use autonomous bike that can be either an electrically-assisted tricycle for passenger commuting or an autonomous carrier for package delivery.

My work

I am primarily in charge of applying multiple local path planners to examine the relationship between the machinery and dynamics of PEV. Another contribution is related to the human-machine interface, including implementing ultra-wideband chips for following function and mechanical eyeballs for improving interactions with pedestrians.

Prototype to production

PEV is originally a research project of the MIT Media Lab. It started from a small shoebox sized wooden prototype and underwent four real-scale prototypes until it became  an actual product, which is now commercialized by Denso Corp.,a Tier-1 automotive supplier.

633037.jpg

PEV v1.4

  • Use standard bicycle components

  • Lightweight yet robust

  • A 250W mid-drive electric motor

  • A 10Ah battery pack that provides 25 miles of travel per charge

  • Top speed of 20 miles per hour

633041.jpg

PEV v2.0

  • Power up to a 350W mid-drive electric motor

  • Air Curtain: the curtain provides cooler air directly to the rider

  • Thermal Storage: the embedded HVAC system keeps the storage cool for package delivery

  • Facial Recognition: The inner camera detects whether the rider is paying attention to the road

Human-machine interface

To address pedestrians' scenario of not knowing the intention of autonomous mobiles, the PEV needs to learn how to interact with the pedestrian seamlessly.

633036.jpg

Mech Eyes v1.0 (prototype)

  • Light indication made by led rings and mirror with three basic modes - idle, pending, and warning

  • Build-in cameras to capture pedestrians' emotions

633040.jpg

Mech Eyes v2.0 (+audio indication)

  • Add Audio indication for warning pedestrians and sending signals if the user is outside the range of 3 meters in the following mode

Simulation

In reviewing different path planners, most planners are designed to fit in some specific mobile platforms. Therefore, I created a new Universal Robot Description Format of the second generation of PEV simulating in virtual environments to quickly verify the efficiency, save time on moving the PEV from places to places and avoid risking damaging the PEV during field testing.

pev_1urdf.png

PEV v1.0

The first version of PEV is only attached to a single CAD model on URDF, and it can only show the position and direction of the PEV but can not present precise potential collision or kinematic performance of the PEV.

ohVuVKI_edited_edited.jpg

PEV v2.0

To have more details on the kinematic and dynamic aspects, I created a new URDF with a separate CAD model for each link. This improvement makes the process of simulation much more accurate and efficient.

Path planner & Following mode

In the scenario of the rider wanting to walk with their personal belongs or goods carried by the PEV, the PEV can be switch to the following mode, also called the donkey mode, utilizing two ultra-wideband anchors embedded to track the rider with an ultra-wideband tag and keep following behind while avoiding obstacles simultaneously.

633363.png

Following mode

planner.png

Local path planner while following

© Vincy Xiao, 2022

  • github
  • LinkedIn
  • 2
bottom of page