Co-Driver is an advanced warning system for safe driving.
Role: UX, UI Designer
Tools: Adobe Illustrator, Photoshop & Premiere Pro - a Sketchbook - Google surveys - Stormboard - Optimal Sort.
Project Length: 3 months
Deliverables: 10 minute presentation and final report.
Responsibility
Define the scoop of the challenge
Develop and perform qualitative research
Define target audience
Build low/high fidelity interactive prototypes
Perform user testing and code results
Define research results into design solution
The Challenge
Assist drivers that need be vigilant of their actions along with multiple types of vehicles all negotiating the same roadway.
Brief
Drivers in densely populated cities have to be aware at all times for pedestrians, bicyclists, and other vehicles that all share the same road space. Non-verbal forms of communication or no communication are the cause of confusions on the roads.
Co-Driver is the answer to the problem of not being aware of all the facets in the vehicles environment with level 1 autonomous driving. Co-Driver helps by giving heads up cues of possible obstacles during the trip.
Case study findings
8 participants were tested in the usability test. 75% said they would use Co-Driver in their vehicles. The feature most wanted by 50% was “assistance with other cars merging into your lane”. The most interesting finding in the data was in the second round of testing 88.9% of the participants said that they do want green light alerts as compared to the first round that 0% said they wanted it. We found that when the question was asked “Do you see yourself using the green light alert” as compared to on the second test “Would you want other drivers to use the green light indicator feature?” more participants selected they would like to have the feature.
Continue below for in-depth processes and recommendations.
Discovery
Observation and Survey were the two research methods used.
fig. 1 Observation notes taken in google doc. click to enlarge.
First research method in solving the problem of driver communication was observations. Seen in figure 1 or here.
Observation of 4 drivers were taken.
3 were aware of the observation.
3 did use signals when merging lanes or turning.
2 were distracted while driving
Sitting at stop lights for too long
Looking at social media
Texting/calls
Using what was observed from the 4 drivers, a survey was developed. Implementing social media and neighborhood forums, as to not use an exhausted university population that may not drive often, a snowball survey sampling method started. Seen in figure 2. The aims of this was to further understand how drivers use the roads and their pain points.
fig.2 Research method two surveys using google forms. click to enlarge.
16 participants responded to the survey.
Participants felt other drivers are:
Impatient
Constantly texting
Lacking with signaling
Not paying attention at lights
Reckless in bad weather conditions
Of the 16 participants who participated in the survey:
9 out of 16 had a fully positive sentiment about tech assist while driving.
6 were open to it but skeptical, worried it could be distracting.
Define and Iteration
Design implications
Based on the results from the observations and survey, we were able to define 3 main issues users have with driving awareness.
Lack of Signaling
Understanding surroundings (both with nearby vehicles and bicyclists)
Distracted driving
fg. 3 Digital post-its in Stormboard. click to enlarge.
An affinity map was built to centralize the data from the surveys and observations. Seen in figure 3.
fig 4. Bicyclist and driver journey maps. click to enlarge.
To build empathy for the users, user journey maps for user-driver and user-bicyclist were builts. Seen in figure 4.
Ideation
Through idea mapping with digital post-it notes and journey mapping, the focus developed.
We decided on creating a product concept called Co-Driver. It is a combination of Augmented Reality and Voice User Interface, allowing the user to choose which they prefer, or both.
There are four main scenarios that Co-Driver looks to solve:
Cars merging into your lane.
Drivers being notified that the light is green.
Notification of nearby bicyclists.
Safely merging into lanes and turning onto a street.
Prototype and Testing
First round of testing
Safety was the top priority, in testing and the goal of the product. We needed to give driving cues while not startling or distracting the driver. Ideation started with inter-group testing and sketching out task scenarios to understand how the cues or alerts would notify the driver. Seen in figure 5 and 6.
“Early internal team usability testing utilizing an ipad was performed to evaluate notifications and visual placement. ”
fig 5. In group testing with ipad of green light progress bar. Axure was used on a ipad to play as the infotainment system using Co-Driver.
fig. 6. Sketches of four different ideas of how the notifications could look.
“User testing had an extra challenge of social distancing during March 2020.”
User testing had an extra challenge of social distancing during the spring of 2020. Brainstorming ways of simulating driving in a safe remote way were challenging but accomplished. A solution to the remote testing was to do the first iteration for tests as a conceptual prototype.
Utilizing google slides as a testing medium with the thought, that most were familiar with google slides at an attempt to remove any extra confusion from the test. The participant were given four scenarios, they would see still images of a driver receiving notifications from Co-Driver and after viewing all scenarios the participants were sent to a questionnaire to give their impressions of the product. From the questionnaire initial information we were looking for was;
Will users understand what Co-Driver is?
Will users want Co-Driver features in their car? What ones that they liked more?
Is there a higher demand for either augmented reality vs. the voice user interface, or will users want both?
Since the participants were going through the prototype remotely, they were given some background on the product concept. Seen in the figure 7 & 8, before starting the scenario prototype, seen in figures 9 and image w/audio 10.
fig. 7 Screen 1 background on concept
fig 8. Screen 2 background on concept
Participants were guided through the scenarios seen in the two examples below, before being sent to the questionnaire.
fig 9. Screen 3 introduction to scenario followed by still image with voice over when clicked. Seen in video to the right.
The test continued with three more scenarios following in the same fashion.
Scenario 2: The light is now green, time to go. If driver didn’t proceed through the intersection a icon of a streetlight was appear on the windshield above the steering wheel with a voice would play “Light is now green”.
Scenario 3: Watch out, there are cyclists next to you. When the vehicle was approaching a cyclists a bike icon would appear on the left side of the windshield. There was no audio for scenario 3 because in heavily populated cities there could be a large amount of cyclists that the driver would encounter. Not wanting to create notification deafness, it was left as visual notification only.
Scenario 4: Looks like you’re wanting to merge. When the driver either turned on the turn signal or started to turn the steering wheel, a yellow swooping arrow would appear on the windshield if there was another vehicle in the path. Once the it was safe to move a voice would say “safe to merge”.
At the end of the scenarios the participants was sent to the questionnaire.
The questionnaire that the participants were sent to after viewing the four scenarios.
The feedback that was given from the participants was,
User's mental model was confused
“How do you install it? Do you need a tech to install it?”
“I thought it was going to be on the infotainment screen.”
General feedback on icon placement and colors
“ I would prefer to have the rear of the car instead of the front.”
“Maybe have the bike icon on the side the bike is on.”
What was interesting about the data from the first round of testing was, in the research phase there was an overwhelming amount stated by the participants about drivers being unaware of a red lights changing to green. Here we see that zero participants selected to have green light indication.
Second round of testing
In the second round of testing, instead of still images the tester would see short video of the product in action. Elements that were updated in the second iteration were,
Icons reflected action happening
Linked colors with real world knowledge
Repositioned alerts
The main questions we were looking to have answered were:
Now that we show what is happening in a video, do users want this product?
What is the a feature users would most want?
When changing the phrasing of the green light indicator are users more receptive?
First screen seen by participants is an introduction to give background before video. Seen to the right.
Followed by a short debrief screen.
Feedback from participants:
Users added they would like emergency vehicle locating similar to bicycle alert.
“Sometimes you can hear the siren but can’t tell if it’s coming from in front of you or from behind you.”
Adding a tone for bicycle alert.
“There’s so many other things to pay attention to when driving, I don’t know how much I’d actually be looking at the heads up display. I feel like I would forget about it after driving for awhile.”
“for me, the voice alerts seem like a distraction. I can see it helpful with green light reminder but merging and car alerts may be unnecessary”
Findings
8 participants were tested in the usability test. 75% said they would use Co-Driver in their vehicles. The feature most wanted by 50% was “assistance with other cars merging into your lane”. The most interesting finding in the data was in the second round of testing 88.9% of the participants said that they do want green light alerts as compared to the first round that 0% said they wanted it. We found that when the question was asked “Do you see yourself using the green light alert” as compared to on the second test “Would you want other drivers to use the green light indicator feature?” more participants selected they would like to have the feature.
For continued iterations:
Haptics in the system. Users that need hearing assistants would have the option of using haptics to accompany visual information.
Mobile devices for cars without heads up displays built in.
Shareable mobile device that goes with the user from car to bicyclists, and to motorcycles.
Research more into feasibility infrared radar, sensors, and video cameras.
Research further into visuals:
Symbiotics
Use of full windshield space in front of drive
When social distancing opens, prototyping in car testing in a safe manner.