Enhancing forklift safety in warehouse operations
I led the design of a blind-spot detection system for forklifts. It improves operator awareness and reduces risks in dynamic warehouses.
Company: Toyota Material Handling Europe (TMHE)
The Result?
A simple 3D display with smart alerts that operators liked, it means less neck strain and more confident driving.
Spoiler: It could save lives and prevent damage at Toyota.
Business impact
Low-cost and flexible implementation
Easy retrofit to existing forklifts
Compatible with current technology (cameras and sensors)
Potential for real-world accident reduction
While lab testing showed promising results like improved confidence in spotting hazards, real-world use could demonstrate actual drops in accidents.
User experience results
Improved confidence
Enhanced hazard detection
"The display is clean and shows distances well."
Helpful without annoyance
Effective audio alerts
"The sound grabbed my attention without stress."
Physical relief
Reduced physical strain
Less head-turning was needed, reducing physical strain for operators.
Low stress levels
Stress levels under 1.3 / 5
Stress levels stayed low during operation with the prototype.
Minimal distractions
Distractions under 2.6 / 5
System maintained focus while providing necessary alerts.
Faster reactions
Enhanced situational awareness
The mix of visual and audio alerts boosted situational awareness and made reacting easier.
Was it easy?
No. Sensor problems in the end and tight deadlines were tough.
What about working in a big organization like Toyota?
I found duplicate researches on things like 360-degree cameras across departments in two offices.
I had to collect and organize them to avoid repeating work.
This also meant aligning engineers and product managers from both offices so everyone knew what the other team had done.
What was the biggest challenge?
Mixing new ideas with rules and costs.
I also had to convince stakeholders to listen to drivers' needs instead of just pushing for innovation. The R&I department focuses on innovative solutions, so they sometimes ignore feasibility or high costs.
What's next?
We handed it over to Toyota's R&D team to build it fully.
Dynamic view expansion
Operators wanted the display to expand in the driving direction, like showing more rear detail when reversing. This gives better focus on key areas, as the opposite side matters less for obstacles.
Before
After
Iterations
In Toyota's lab, we set up the prototype on a real forklift. Sensors had issues during testing, so we used a Wizard of Oz method—manual triggers to simulate alerts. I then demoed the updated version to stakeholders. This led to useful feedback for improvements.
How it works
The blind-spot assistant uses sensor data to show a top-down 3D forklift model on a display. It uses colored zones and dots for obstacles. Sound alerts turn on for closer dangers. We used visual and audio cues to lower cognitive load. Here is how it works in different cases:
No obstacle state
When nothing is nearby, the display shows a plain circle around the forklift. This keeps it simple and helps the driver stay focused.
Distant obstacle state
For objects between 0.5 and 1.5 meters away, a yellow area lights up in that direction. A steady yellow dot shows the spot. It just reminds the driver to be careful, no rush.
Close obstacle state
If an obstacle is under 0.5 meters away, a red area and blinking red dot warn of danger. Sound alerts get louder to make the driver react fast, like stopping right away.
Prototyping
I developed four prototype versions and collaborated with engineers to build and test them on a forklift in Toyota's lab.
Why a 3D model display?
The 3D model display won because it uses existing sensors for quick rollout and scalability across forklift models. It works reliably in dusty or low-light warehouses, and we can always add cameras later for better accuracy, like cutting down false alarms.
For operators, it gives a clear view of hazard locations without flooding them with details. It’s just enough for them to react confidently.
Financially, it stays affordable by relying on existing sensors or lower-grade cameras, since we process images behind the scenes instead of streaming high-quality video. It's technically solid, budget-friendly, and trusted by drivers for safer operations.
Selecting the solution
The workshop highlighted a 3D forklift model display with sound alerts as the top pick. It stood out for its simplicity and minimal distraction, allowing operators to stay focused without overwhelming them.
Generating Ideas
The focus group insights showed camera systems had issues like high costs, maintenance, and mast obstructions. This prompted us to brainstorm solutions beyond cameras, from haptic feedback to wearable alerts. (Details are confidential, but I can discuss them in person.) To prepare for the workshop, I created storyboards to illustrate these in warehouse scenarios.
Aligning stakeholders
I observed separate research efforts across departments on cameras and sensors, with limited cross-team awareness. Engineers focused on technology, product managers on costs, often overlooking operator needs.
To bridge this, I organized a workshop with operators, logistic managers, program managers, product managers, and engineers. We evaluated ideas based on technical feasibility, financial viability, and operator desirability. This built consensus, as the votes showed.
Still scrolling?
Awesome. Let's dive into how we turned warehouse chaos into safer, smoother operations.
My goal?
Design an intuitive blind-spot system that improves situational awareness, reduces cognitive and physical strain, scales across forklift models for business value, and delivers cost-effective reliability in warehouse conditions.
First thing first
Uncovering the real problems
Through targeted interviews and focus groups, I identified two categories of challenges.
Category 1
User experience challenges
From interviews with forklift operators and warehouse supervisors.
Harsh environment
Poor lighting, loud noise, and tight spaces make it harder to detect objects.
Mental overload
80% of errors come from poor situational awareness, especially under stress and time pressure.
Physical strain & fatigue
Frequent head turns to check behind cause fatigue and discomfort.
Visibility incidents
These lead to pedestrian injuries, near-misses, and collisions with racks or pallets.
User Experience Requirements
Screen size and placement
An iPad-sized scree, near the mirror or right dashboard side.
Frequent alerts
Avoid alert overload to prevent drivers from ignoring warnings.
Obstacle types
No need to differentiate; all hazards require equal attention.
Distraction concerns
Multiple alerts or real-time camera feeds can divert attention.
Category 2
Technical challenges
From focus groups with ADAS engineers and product managers.
Assistance vs. safety feature
Operators must know assistive tech differs from safety features. They shouldn't fully rely on the system and must stay aware of surroundings. Based on regulations, they remain responsible for any incidents.
Maintenance and retrofitting
Vibrations and dust shorten camera lifespan, needing durable parts and wiring cameras on older forklifts is tough and limits scalability across different forklift types.
Camera limitations
Cameras struggle in low light, dust, or temperature shifts between warm and cold storage. This can reduce image quality and reliability.
Camera obstruction
In reach trucks, the project's focus, a large moving mast on the right blocks camera views when extended.
High processing costs
Camera systems need costly hardware for high-bandwidth real time image processing.
False alerts
Sensors might mistake racks for hazards. Cameras with image processing could help reduce this.
Prioritizing features
Next, I compared Toyota's safety features with those of competitors and car technologies, which revealed gaps in solutions. To address these, I led a workshop with product managers, engineers, and program managers to review potential features and set priorities.
The top priority was a 360-degree camera with blind-spot assistant. This system allows for future expansions to other features.
Blind spot challenge
Blind spots cause 80% of forklift accidents due to limited visibility.
Operators often turned their bodies to check hidden areas, risking near-misses in busy spaces.
Getting started
Planning the safety strategy
I began with desk research and on-site observations at TMHE's factory to pinpoint core issues. Blind spots stood out as a major concern.
My role
UX Designer & Researcher
As the main designer and strategist on this project, I led the end-to-end process, from research to prototyping,. Reporting to senior stakeholders in the R&I department.
What are we trying to accomplish?
Toyota was behind competitors in advanced safety features. The goal was to increase operational safety in warehouses by adapting Advanced Driver Assistance Systems (ADAS) from cars, such as blind-spot warnings, collision alerts, or lane assistance, to forklifts.
Enhancing forklift safety in warehouse operations
Enhancing forklift safety in warehouse operations
I led the design of a blind-spot detection system for forklifts. It improves operator awareness and reduces risks in dynamic warehouses.
I led the design of a blind-spot detection system for forklifts. It improves operator awareness and reduces risks in dynamic warehouses.
Company: Toyota Material Handling Europe (TMHE)
Company: Toyota Material Handling Europe (TMHE)
The Result?
A clean 3D display with smart alerts that operators appreciated. It leads to less physical strain and more secure driving.
The Result?
A clean 3D display with smart alerts that operators appreciated. It leads to less physical strain and more secure driving.
Spoiler: It shows promise for saving lives and minimizing damage at Toyota.
Spoiler: It shows promise for saving lives and minimizing damage at Toyota.
What are we trying to accomplish?
What are we trying to accomplish?
Toyota was behind competitors in advanced safety features. The goal was to increase operational safety in warehouses by adapting Advanced Driver Assistance Systems (ADAS) from cars, such as blind-spot warnings, collision alerts, or lane assistance, to forklifts.
Toyota was behind competitors in advanced safety features. The goal was to increase operational safety in warehouses by adapting Advanced Driver Assistance Systems (ADAS) from cars, such as blind-spot warnings, collision alerts, or lane assistance, to forklifts.
My role
My role
UX Designer & Researcher
UX Designer & Researcher
As the main designer and strategist on this project, I led the end-to-end process, from research to prototyping,. Reporting to senior stakeholders in the R&I department.
As the main designer and strategist on this project, I led the end-to-end process, from research to prototyping,. Reporting to senior stakeholders in the R&I department.
Getting started
Getting started
Planning the safety strategy
Planning the safety strategy
I began with desk research and on-site observations at TMHE's factory to pinpoint core issues. Blind spots stood out as a major concern.
I began with desk research and on-site observations at TMHE's factory to pinpoint core issues. Blind spots stood out as a major concern.
Blind spot challenge
Blind spot challenge
Blind spots cause 80% of forklift accidents due to limited visibility.
Operators often turned their bodies to check hidden areas, risking near-misses in busy spaces.
Blind spots cause 80% of forklift accidents due to limited visibility.
Operators often turned their bodies to check hidden areas, risking near-misses in busy spaces.
Prioritizing features
Prioritizing features
Next, I compared Toyota's safety features with those of competitors and car technologies, which revealed gaps in solutions. To address these, I led a workshop with product managers, engineers, and program managers to review potential features and set priorities.
Next, I compared Toyota's safety features with those of competitors and car technologies, which revealed gaps in solutions. To address these, I led a workshop with product managers, engineers, and program managers to review potential features and set priorities.
The top priority was a 360-degree camera with blind-spot assistant. This system allows for future expansions to other features.
The top priority was a 360-degree camera with blind-spot assistant. This system allows for future expansions to other features.
First thing first
First thing first
Uncovering the real problems
Uncovering the real problems
Through targeted interviews and focus groups, I identified two categories of challenges.
Through targeted interviews and focus groups, I identified two categories of challenges.
Category 1
User experience challenges
User experience challenges
From interviews with forklift operators and warehouse supervisors.
Harsh environment
Poor lighting, loud noise, and tight spaces make it harder to detect objects.
Poor lighting, loud noise, and tight spaces make it harder to detect objects.
Mental overload
80% of errors come from poor situational awareness, especially under stress and time pressure.
80% of errors come from poor situational awareness, especially under stress and time pressure.
Physical strain & fatigue
Frequent head turns to check behind cause fatigue and discomfort.
Frequent head turns to check behind cause fatigue and discomfort.
Visibility incidents
These lead to pedestrian injuries, near-misses, and collisions with racks or pallets.
These lead to pedestrian injuries, near-misses, and collisions with racks or pallets.
User Experience Requirements
User Experience Requirements
Screen size and placement
An iPad-sized scree, near the mirror or right dashboard side.
An iPad-sized scree, near the mirror or right dashboard side.
Frequent alerts
Avoid alert overload to prevent drivers from ignoring warnings.
Avoid alert overload to prevent drivers from ignoring warnings.
Obstacle types
No need to differentiate; all hazards require equal attention.
No need to differentiate; all hazards require equal attention.
Distraction concerns
Multiple alerts or real-time camera feeds can divert attention.
Multiple alerts or real-time camera feeds can divert attention.
Category 2
Technical challenges
Technical challenges
From focus groups with ADAS engineers and product managers.
Assistance vs. safety feature
Operators must know assistive tech differs from safety features. They shouldn't fully rely on the system and must stay aware of surroundings. Based on regulations, they remain responsible for any incidents.
Maintenance and retrofitting
Vibrations and dust shorten camera lifespan, needing durable parts and wiring cameras on older forklifts is tough and limits scalability across different forklift types.
Camera limitations
Cameras struggle in low light, dust, or temperature shifts between warm and cold storage. This can reduce image quality and reliability.
Camera obstruction
In reach trucks, the project's focus, a large moving mast on the right blocks camera views when extended.
High processing costs
Camera systems need costly hardware for high-bandwidth real time image processing.
False alerts
Sensors might mistake racks for hazards. Cameras with image processing could help reduce this.
My goal?
My goal?
Design an intuitive blind-spot system that improves situational awareness, reduces cognitive and physical strain, scales across forklift models for business value, and delivers cost-effective reliability in warehouse conditions.
Design an intuitive blind-spot system that improves situational awareness, reduces cognitive and physical strain, scales across forklift models for business value, and delivers cost-effective reliability in warehouse conditions.
Still scrolling?
Still scrolling?
Awesome. Let's dive into how we turned warehouse chaos into safer, smoother operations.
Awesome. Let's dive into how we turned warehouse chaos into safer, smoother operations.
Aligning stakeholders
Aligning stakeholders
I observed separate research efforts across departments on cameras and sensors, with limited cross-team awareness. Engineers focused on technology, product managers on costs, often overlooking operator needs.
I observed separate research efforts across departments on cameras and sensors, with limited cross-team awareness. Engineers focused on technology, product managers on costs, often overlooking operator needs.
To bridge this, I organized a workshop with operators, logistic managers, program managers, product managers, and engineers. We evaluated ideas based on technical feasibility, financial viability, and operator desirability. This built consensus, as the votes showed.
To bridge this, I organized a workshop with operators, logistic managers, program managers, product managers, and engineers. We evaluated ideas based on technical feasibility, financial viability, and operator desirability. This built consensus, as the votes showed.
Generating Ideas
Generating Ideas
The focus group insights showed camera systems had issues like high costs, maintenance, and mast obstructions. This prompted us to brainstorm solutions beyond cameras, from haptic feedback to wearable alerts. (Details are confidential, but I can discuss them in person.) To prepare for the workshop, I created storyboards to illustrate these in warehouse scenarios.
The focus group insights showed camera systems had issues like high costs, maintenance, and mast obstructions. This prompted us to brainstorm solutions beyond cameras, from haptic feedback to wearable alerts. (Details are confidential, but I can discuss them in person.) To prepare for the workshop, I created storyboards to illustrate these in warehouse scenarios.
Selecting the solution
Selecting the solution
The workshop highlighted a 3D forklift model display with sound alerts as the top pick. It stood out for its simplicity and minimal distraction, allowing operators to stay focused without overwhelming them.
The workshop highlighted a 3D forklift model display with sound alerts as the top pick. It stood out for its simplicity and minimal distraction, allowing operators to stay focused without overwhelming them.
Why a 3D model display?
Why a 3D model display?
The 3D model display won because it uses existing sensors for quick rollout and scalability across forklift models. It works reliably in dusty or low-light warehouses, and we can always add cameras later for better accuracy, like cutting down false alarms.
The 3D model display won because it uses existing sensors for quick rollout and scalability across forklift models. It works reliably in dusty or low-light warehouses, and we can always add cameras later for better accuracy, like cutting down false alarms.
For operators, it gives a clear view of hazard locations without flooding them with details. It’s just enough for them to react confidently.
Financially, it stays affordable by relying on existing sensors or lower-grade cameras, since we process images behind the scenes instead of streaming high-quality video. It's technically solid, budget-friendly, and trusted by drivers for safer operations.
For operators, it gives a clear view of hazard locations without flooding them with details. It’s just enough for them to react confidently.
Financially, it stays affordable by relying on existing sensors or lower-grade cameras, since we process images behind the scenes instead of streaming high-quality video. It's technically solid, budget-friendly, and trusted by drivers for safer operations.
Prototyping
Prototyping
I developed four prototype versions and collaborated with engineers to build and test them on a forklift in Toyota's lab.
I developed four prototype versions and collaborated with engineers to build and test them on a forklift in Toyota's lab.
How it works
The blind-spot assistant takes sensor data and shows a top-down 3D model of the forklift on the display. Hazards appear as colored zones and dots, with sound alerts kicking in when something gets too close. By combining visual and audio cues, it taps into multiple senses and lowers cognitive load.
No obstacle state
No obstacle state
When nothing's nearby, the display shows a plain circle around the forklift. This keeps it simple and helps the driver stay focused.
Distant obstacle state
Distant obstacle state
For objects 0.5–1.5m away, a yellow zone lights up in that direction. A steady yellow dot shows the spot. It just reminds the driver to be careful, no rush.
Close obstacle state
Close obstacle state
If something is under 0.5m, the zone turns red and the dot blinks. Sound alerts get louder to push for quick action, like stopping immediately.
Iterations
Iterations
In Toyota's lab, we set up the prototype on a real forklift. Sensors had issues during testing, so we used a Wizard of Oz method, manual triggers to simulate alerts. I then demoed the updated version to stakeholders, which brought in valuable feedback for improvements.
In Toyota's lab, we set up the prototype on a real forklift. Sensors had issues during testing, so we used a Wizard of Oz method, manual triggers to simulate alerts. I then demoed the updated version to stakeholders, which brought in valuable feedback for improvements.
Dynamic view expansion
Dynamic view expansion
Operators wanted the display to stretch in the driving direction. This gave them more depth perception and helped them locate obstacles more precisely, while filtering out less relevant areas behind them.
Before
Before
After
After
User experience results
User experience results
Improved confidence
Enhanced hazard detection
"The display is clean and shows distances well."
Helpful without annoyance
Effective audio alerts
"The sound grabbed my attention without stress."
Physical relief
Reduced physical strain
Less head-turning was needed, reducing physical strain for operators.
Low stress levels
Stress levels under 1.3 / 5
Stress levels stayed low during operation with the prototype.
Minimal distractions
Distractions under 2.6 / 5
System maintained focus while providing necessary alerts.
Faster reactions
Enhanced situational awareness
The mix of visual and audio alerts boosted situational awareness and made reacting easier.
Business impact
Low-cost and flexible implementation
Easy retrofit to existing forklifts
Compatible with current technology (cameras and sensors)
Potential for real-world accident reduction
While lab testing showed promising results like improved confidence in spotting hazards, real-world use could demonstrate actual drops in accidents.
While lab testing showed promising results like improved confidence in spotting hazards, real-world use could demonstrate actual drops in accidents.
Was it easy?
Was it easy?
No. Sensor problems in the end and tight deadlines were tough.
No. Sensor problems in the end and tight deadlines were tough.
What about working in a big organization like Toyota?
What about working in a big organization like Toyota?
I found duplicate researches on things like 360-degree cameras across departments in two offices.
I found duplicate researches on things like 360-degree cameras across departments in two offices.
I had to collect and organize them to avoid repeating work.
I had to collect and organize them to avoid repeating work.
This also meant aligning engineers and product managers from both offices so everyone knew what the other team had done.
This also meant aligning engineers and product managers from both offices so everyone knew what the other team had done.
What was the biggest challenge?
What was the biggest challenge?
Mixing new ideas with rules and costs.
Mixing new ideas with rules and costs.
I also had to convince stakeholders to listen to drivers' needs instead of just pushing for innovation. The R&I department focuses on innovative solutions, so they sometimes ignore feasibility or high costs.
I also had to convince stakeholders to listen to drivers' needs instead of just pushing for innovation. The R&I department focuses on innovative solutions, so they sometimes ignore feasibility or high costs.
What's next?
What's next?
We handed it over to Toyota's R&D team to build it fully.
We handed it over to Toyota's R&D team to build it fully.