The Honda EV project was a study to test how changes in interface designs could help people trust an electric vehicle more when presented with a range anxiety situation. I designed two different versions of a possible new interface for an EV. We had a 2x2 matrix study design where participants were in one of four categories: full charge exact range display, low charge exact range display, full charge inexact range display, and low charge exact range display.
Participants drove the car for approximately 30 minutes. In the low charge conditions, the range display led participants to believe that they did not have enough charge to reach their destination. We then surveyed participants on a variety of aspects including their trust in the vehicle and comfort with the interface.
In our results we found that even though people liked the exact range display better, they trusted the car more when shown the inexact display. We believe we saw these results because though the inexact range display might not be very practical, the exact range display can be misleading, particularly when driving conditions change.
We used an Android tablet to display our interface prototype. The tablet fit easily in front of the original Honda dashboard display.
One of my main contributions to the study was the interface design. First, I researched the type of information that EVs can display to drivers based on what is available in the market. I did a comprehensive literature review to define the types of interfaces and methods of testing that had been used in the past in human-vehicle interaction. The biggest challenge for me was to design an interface that was both relevant to current and future EV drivers and testable in a scientific setting.
I sketched numerous possible interfaces, first simply using pen and marker and then using Photoshop and PowerPoint. I iterated on the design many times after receiving feedback from our PI and partners at Honda. Below I included a sample of the interface iterations and my decisions behind them.
This display involves incorporating state of charge, miles driven on that charge, and range to depict to the user how these pieces of information interact. My goal for these interfaces was to display to the user that state of charge and range are linked.
The display makes it clear to the driver that driving behavior and conditions affect range, and lets a driver become more in-tune with the mechanisms of the car. However, the sketch still displays a numberical number for range, and could possibly be providing the user with too much information than to be useful when driving.
This concept sketch displays the state of charge and range in a single graphic. This sketch makes it clear to the user that SOC and range are linked, and it also displays a gradient, rather than an exact number, for range. However, the graphic could be hard to read as the car's charge becomes low, and once again, could be displaying too much information than to be useful.
This interface iteration (made in Photoshop) displays a variety of information useful to the driver: speed, state of charge, range, and power consumed versus time. I placed the state of charge and range displays on top of each other so that the driver can see how state of charge and range drop at approximately the same rate. I included the power consumption versus time graph to show the driver how his or her driving behavior (among other factors) affects the vehicle's power consumption.
I designed this display to include the maximum amount of information to the driver in a helpful manner. Though seeing power consumption versus time might be helpful to see, in the end we needed to test only small changes in an interface display, and so I simplified the design to come to the end result.
The final interface for the inexact range conditions. We showed participants their speed, state of charge, and range, and nothing else.
The final interface for the exact range conditions.
Above are the two versions of the interface that we used for the experiment. I changed the speedometer to a dial speedometer, as people are typically more familiar with a dial than a digital readout. I also changed the state of charge display to mimic that of a gas display in a typical car. The state of charge display deliberately only has full, empty, and the halfway point marked, as I wanted to encourage people to focus on the range display.
Above is the poster that I presented at Stanford's mechanical engineering Undergraduate Research Summer Poster Presentation.
