Designing an In-Cab Alert System for the 2025 Mercedes Benz “Truck of the Future”
The Mercedes Benz semi-truck of 2025 will only be autonomous on the highway and only in optimal road conditions. A driver will still be required to be in the cab. Our team was faced with answering the challenging question of how you would alert drivers when they were needed to take control of the vehicle?
Two week project for Interaction Design Practice
We designed an alert system for autonomous semi-trucks in 2025 that alerts drivers when they need to take control of the vehicle. We used voice and light alerts in order to grab the drivers attention when they were needed. We also designed a smart system that would safely pull over if the driver did not respond to the alerts.
WHAT I LEARNED
Stay on the same page as your team by writing down where you’ve been. My team and I did a lot of research for this project and explored many different design paths. We started each meeting by writing down where we had been and the design decisions we had made. By writing it out on a whiteboard we were able to quickly discuss what direction we wanted to pursue next and remember what decisions had influenced us up until that point.
- Write down your assumptions about your users and then question them. Everyone has assumptions. It’s important to know your assumptions and then question them by going out and talking to people and doing in-depth research.
- Designing for an experience is difficult. It’s one thing to design an interface, its entirely different challenge to design a experience where you have voice alerts and lights working together to achieve a goal. It completely changes the way you sketch, prototype, and user test your design.
- Get outside your design studio! For this project we traveled to truck stops and distribution warehouses where we got to spend afternoons talking to truck drivers, mechanics, warehouse workers, and truck salesmen. We were able to gain incredible insights that wouldn’t have happened if we would have just stayed inside doing “research” on our computers all day.
PROBLEM TO SOLUTION - THE PROCESS
My team was tasked with designing a feature that would address a future ethical, safety, or trust issue brought on by the advent of autonomous vehicles. Our client was Mercedes Benz. The client brief was left open-ended. It was our responsibility as a team to find a core problem that could be solved within the timeline of the project.
Choosing a Mercedes Benz vehicle to focus on
It became apparent early on that we needed to make quick team decisions on what Mercedes Benz vehicle and autonomous problem we wanted to tackle. We did research online and discovered that Mercedes Benz manufacturers all kinds of different vehicles. This gave us a lot of options to choose from.
We wrote out the list of vehicles on a whiteboard and, after a bit of back and forth, decided to focus on semi-trucks. We discovered that Mercedes Benz had already designed and developed an autonomous “semi-truck of the future” for 2025. After exploring what fully autonomous trucks with no drivers would look like, we decided it would be best to use the context of 2025 and the Mercedes Benz semi-truck of the future as our focus for the project.
Mercedes Benz 2025 "Truck of the Future"
Through our research we discovered that companies are approaching self-driving vehicles differently. Some companies like Google are skipping over semi-autonomy where the human has the ability to take control of the wheel from the computer in various situations. Other companies like Tesla and Mercedes Benz are creating vehicles that allow the driver to drive as wanted or needed.
The Mercedes Benz semi-truck of the future has a feature called “Highway Pilot” that allows the driver to take their hands off the wheel and relax while the computer on board takes control and drives autonomously down the highway. In 2025, a truck driver will still be needed in the cab of the truck in order to take care of problems as they arise. Most importantly a human driver is still needed to do complex driving tasks such driving through inclement weather, heavy construction, cities/roads off the highway, and backing up the trailer to load and unload goods.
Throughout the project, we hit the road and met the people and companies behind the $700 billion dollar trucking industry in America. We spent a lot of our time during the project talking to people involved in the trucking industry. Our goal was to soak up as much information as we could about the people and companies who move goods across the country everyday.
Visiting a truck stop, mechanic shop, and truck sales lot
Being based out of Bloomington, Indiana we decided to drive up to Indianapolis which is a major trucking hub in the United States. We met with truck drivers and a salesman that provided us with valuable information. We were lucky to meet amazing people there that were more than willing to share their experience and let us ask them questions.
Some valuable things we learned were:
- Truck drivers do way more than sit and drive a truck all day
- The trucking industry is absolutely crucial to the American economy
- Truck drivers are on tight schedules which makes it hard for them to follow all the safety rules and regulations set by the Department of Transportation (DOT)
Truck inspections are a tedious process that take up precious time
Taking a tour through a truck distribution center
During the project we also made contact with a truck distribution center that allowed us to tour their warehouse and talk to their drivers and logistics managers. Throughout the project we continued to keep in touch with these folks as our design evolved and changed over time. They made time for us to continue coming in to their office to discuss our sketches and they even let us test our prototype on them. Thanks HD Supply!
Some valuable things we learned were:
- Truck drivers pre-trip preparations are a complex process
- Loading and unloading the trailers takes extreme amounts of organization
- Backing up a semi-truck to a unloading dock takes a lot of practice
FINDING A CORE PROBLEM
We explored many different areas in regards to semi-trucks and autonomous vehicles before landing on our final design. We would explore a certain idea that we had only to realize that it wasn’t the right direction for us to head in after conversations with experts in the industry, sketching sessions, and doing research online.
Some of the areas we explored were:
- Fully autonomous vs semi-autonomous trucks —Early on in the project we wanted to explore what the world would look like for a fully autonomous semi-truck that would navigate the roads without any human drivers in the cab. However, it became clear to us through our research that semi-trucks will be one of the last vehicles to drive completely autonomous due to the amount of work truck drivers do besides driving the truck on the road. We decided if would be beneficial to focus on semi-autonomous trucks and design in the context of 2025 and the Mercedes Benz truck of the future.
- Truck inspections —We then began to focus on one of the problems we heard a lot while talking to drivers. Truck inspections take forever! Drivers are required by the DOT to do safety inspections of their truck and load before driving each load. Our goal by exploring truck inspections was to see if we could find a way to make the process of inspecting the truck easier before you go out on the road. We ideated how we could automate this process. However, we eventually discovered there were already technologies trying to automate this time consuming process by adding sensors all over the truck. We decided to move on.
- Transfer of control from driver to autonomous technology — Finally, we shifted our focus towards the safety issues involved with a semi-truck that can drive autonomously in optimal road conditions yet still needs a human when exiting the highway or in rough weather and road conditions (snow, heavy rain, construction, etc).
We really benefited from doing a design exercise called, “Golden Questions”. The whole goal is to ask as many questions as you can about a specific topic you are exploring. Once you ask them you then ask sub questions based off the first questions you asked. We did this exercise each time as we continued to constrain deeper into a new idea or area. Even though we wouldn't be able to address all the questions it helped us flush out our design and ask the tough questions amongst ourselves before we presented our final design.
To drill down further into the transfer of control we started asking questions like:
- How would the driver give control to the Highway Pilot technology and feel confident that it would safely drive the truck?
- How would the driver take control back from Highway Pilot?
- What happens if another vehicle crashes into the truck while Highway Pilot is driving?
- How would the driver be alerted he or she is needed to take control?
Eventually we honed in on a core set of problems that we believed was absolutely crucial to the safety of the driver and the other cars on the road before the Mercedes Benz truck of the future could safely take the road in 2025.
How do you alert the truck driver when they need to take back driving control from the autonomous system? 🔔
What happens if the driver turns on the autonomous feature and then falls asleep in the back cabin? 💤
After we had shown early sketch ideas to our friends at HD Supply we felt confident enough in our idea to invest time in creating a more realistic prototype.
It was an interesting challenge creating a prototype for an experience. We recorded 8 different voice alerts to test on our users all with subtle differences in tone and information given. To be consistent we used an alert notification that severe weather was approaching and the driver was needed to take control of the wheel.
Testing our prototype was really fun! We reserved a room nearby a cafe where we knew we could get a consistent and random flow of people coming by. We were able to get six people to test our prototype.
The main questions we wanted to answer during user testing was:
a) Would the user hear the voice alert and understand what it said?
b) Would the lights and audio in the cab be enough to grab their attention?
c) How fast would they react to the alert and move back to the drivers seat?
Some key takeaways we had from the usability test was:
- It’s important to keep voice alerts simple and straight to the point
- The volume of the alert matters
- Users typically have a voice alert preference (male vs female)
- Differing voice tones were interpreted as either emotional or authoritative depending on the gender of the voice
- Multiple alarms at decreasing intervals are more effective than one alarm
- Lights can be effective at grabbing the users attention however too many lights flickering is unnecessary
Once we had tested users with our truck cabin alert prototype, we iterated based off their feedback and the problems we noticed when watching them interact with our design.
After our usability testing we made final decisions on our design based on the feedback we received. We chose to have the final alert recorded with a female voice in a calm tone. We also decided against integrating a voice tone change with the recording depending on the time the message is presented. The lighting remained a crucial part of our design as it received positive feedback. However, we would use lights turning on and off sparingly to avoid annoying the user with constant flickers.
FINAL DESIGN AND FEATURES
Our final design addresses the problem of how and when to alert the driver they are needed to take control of the vehicle. We addressed this problem by designing a system that would automatically inform the driver when they were needed by voice and light alerts. We constrained our design to the most extreme safety issue we discovered:
What we did not design:
How the driver would set the Highway Pilot feature
- How the driver transitions after taking back control from the computer
- How the driver would be able to avoid an instant collision ahead
Voice and Light alerts
The meat of our design is the voice and light system that turns on to notify the driver that he/she is needed to take control of the wheel. The system triggers the alerts to go off starting at 40 minute until the driver is needed until there is 1 minute left until Highway Pilot cannot continue driving. Once the time gets down to 5 minutes left the alert frequency picks up in hopes of grabbing the distracted truckers attention.
The onboard system constantly tracks the weather ahead on the route while Highway Pilot is engaged. This allows the driver to have peace of mind that the system will alert him/her if their assistance is needed.
Heat map sensors
Heat map sensors allows the system to determine whether or not the driver has moved since the alerts have started going off. If the sensors determine that the driver is still unresponsive the frequency and sound level of the alerts continue to increase.
Safety pullover spots
In the case of an emergency where the driver does not not take back control from Highway Pilot and the Highway Pilot can no longer safely operate the driver system that we designed automatically pulls over to a designated safety spot on the side of the road. From there Highway Autopilot notifies emergency responders and the trucking company. The Highway Autopilot system automatically determines the nearest safety pullover spot available based on the Google Maps API.
We presented our final design to a room full of first and second year designers. Our design was well received and we were able to get a lot of great critique in order to improve our design. Unfortunately, with the next project right around the corner we weren't able to make changes and improve our design.
Every grad school project ends before you get to explore more areas and refine things as much as you’d like. If we had more time we would explore the following areas:
- Perform a usability test with our design on truck drivers inside their real truck cabins while another truck driver drives and the other one is able to fall asleep
- Explore how we could ensure the driver was awake and aware before taking control (maybe the driver has to answer some basic questions, maybe they have an eye scanner, maybe it requires him to drink coffee, idk)
- Test different voices and customized alerts