Developing a Rear-Vision System for a Bicycle

Real-Time-Rear-Vision-System

Just over a year ago my friend was struck from behind by a car while riding with the local bicycle club. He sustained numerous physical injuries and despite a full recovery he hasn’t touched his bike due to the psychological trauma.

I love biking but the incident made me hesitant to ride on busy roads. Rather than letting blind chance dictate my safety I decided to develop a device to see the entire road behind me. Having a real-time view of the road behind would let me react to the unpredictable behavior of distracted drivers.

Follow my journey from initial concept to fully functional system as I navigate the world of user testing, rapid prototyping, and iterative design.

The Initial Concept

The idea was to mount a rear facing camera on my bike and wirelessly transmit real-time video footage through Bluetooth to my cellphone mounted on the handlebar.

To quickly test the concept I recorded rear-view video footage on my cellphone and then mounted the phone on the handlebar. I immediately became aware of a few issues with the first design concept:

initial-concept

  1. Glare while viewing the video footage (even with anti-glare film applied to the screen)
  2. Screen seemed too small to see approaching vehicle from a distance
  3. Position of the screen distracted from maintaining eye contact on the road ahead
  4. Viewing angle of phone camera too narrow to see cars alongside
  5. Horizontally ‘mirrored’ video footage

Screen Size Testing

Having realized that the phone screen might be too small to see approaching vehicles from a distance I decided to test a variety of screen sizes to identify the optimal trade-off between visibility and practicality.

I mounted a webcam to the back of my seat post and secured a large tablet to my handlebar.

screen-size-testing

I then created 4 different cut out templates that could be quickly overlay on the tablet to simulate different screen sizes while riding.

screen-size-testing2

Testing revealed that a 2.5” diagonal screen was too small, 3.5” only worked in close proximity, and 5” took up too much handlebar real estate. A 4.3” screen offered the best trade-off between good vehicle visibility and handlebar space.

Screen size testing also gave me insight into a few other issues:

  1. The webcam stopped providing a live video feed while riding and remained on a static image. It took me several seconds to realize that I was no longer seeing real-time video. Lesson learned: the final design would need an onscreen indicator to show when video is real-time streaming or connection is lost
  2. The tablet experienced instances of lag while riding, resulting in delayed video feed
  3. The webcam would meter for the entire image causing the road surface portion of the video to become too dark during sunny riding conditions
  4. Glare was once again a major problem

Design Decisions

In addition to the issues I observed during screen size testing there were a few other design decisions that needed to be made. I had to decide on either a wireless or cable connected design. A wireless design would make for easier installation, however; it would require two separate battery packs – one for the camera and the other for the screen. In addition a wireless connection could cause video sync issues if the Bluetooth link was interrupted.

I also had to decide whether the design would use a cellphone or custom display. A phone would be more convenient to use, however; the phone screen would experience glare, was too small to see vehicles at a distance, and the constant video feed would drain the phones battery.

I decided on a cable connected design with a custom handlebar mounted display. The cable would ensure uninterrupted video feed and a custom display could be designed to reduce glare, have a dedicated video processor, and provide the optimal mounting position.

Electronic Hardware Specs

I compiled a block diagram outlining the technical requirements for the camera and custom display. The camera unit would need a 1/3” CMOS digital image sensor and a digital video processor coupled to a wide angle lens. The display unit would need a thin film transistor transflective LCD color screen with optical bonding and anti-reflective outer glass to allow for use in direct sunlight.

The camera and display would be connected using a microUSB cable and powered by a 3.7V battery. A microSD card slot in the display would provide video footage backup in case of an accident.

block-diagram

I contacted numerous manufacturers until I found a supplier that stocked a camera and screen that met my technical specifications. I worked with the manufacturer to modify the display circuit board to include custom connectors, a detached power supply and proprietary firmware to process the video feed. The circuit board for the camera was off the shelf but I tweaked the digital video processor to perform some graphic manipulations to meet the constraints of the display.

Electronic Hardware Testing

While waiting for the parts to arrive I dedicated my time to designing a temporary enclosure to mount the screen hardware for initial performance testing. Being a huge fan of open source I used FreeCAD to draft a simple mounting frame and used my RepRap to 3D print a prototype.

Once the electronic hardware arrived I mounted the screen and checked performance in direct sunlight and low-light conditions. I configured the brightness and contrast for each scenario and tested the viewing angle to ensure good visibility. I also tested the battery life under different scenarios including maximum screen brightness.

electronic-hardware-testing

I tested different lengths of cable between the camera and the display to determine if there was any delay or degradation in video quality. Even with a 28” cable between the CMOS sensor and the digital video processor there was no loss in performance. Once I was satisfied with the performance of the electronics I started designing the display enclosure, camera housing, and mounting hardware.

Display Enclosure Design

I wanted to design the display enclosure as thin as possible to make it lightweight and aerodynamic, however; I was constrained by the thickness of the printed circuit board (PCB) and the LCD screen. I designed the enclosure in three separate components:

  1. Top plate to mount the LCD screen
  2. Thin profile chassis to secure the PCB and hold the LCD screen against the enclosure
  3. Back plate with an integrated quarter-turn connector to allow for mating to a quarter-turn handlebar mounting system.

I tried numerous configurations and finally settled on a two-tone color to visually shrink the enclosure.

FreeCAD-Full-Enclosure-PCB-Mount

I modeled the display in FreeCAD and exported the mesh objects to Blender3D for rendering in LuxRender so that I could preview the display and assess its proportions.

luxrender-enclosure

Once I was satisfied with the design I triple-checked my dimensions and exported the model to Slic3r for 3D printing. Within a few hours I had my high resolution 3D printed parts in hand and it was time to assemble the display.

The LCD screen was inserted into the top plate offset and then secured into place by the PCB chassis, with the PCB mounted upwards on the back of the chassis. Cables were routed through the opening in the back plate and the display was secured with four corner fasteners.

3d-printed-display

3d-printed-display2

3d-printed-display3

Camera Enclosure Design

Having completed the display it was time to design the seat-post mounted rear-view camera. I wanted to create the smallest possible camera housing so I separated the small lens and CMOS sensor from the larger digital video processor. I used FreeCAD to model a two piece enclosure that secured the lens and CMOS sensor between two plates and then routed a ribbon cable to the digital video processor mounted at the display. The camera housing was attached to the seat-post by a swiveling clamp with integrated hex nut to allow for level adjustment.

FreeCAD-Camera-Mount-together-complete

3D-printed-camera-enclosure

Mounting Hardware Design

The last component that I needed to design was the handlebar mounting system. The design needed to be robust, easily adjustable, and allow the display to be quickly removed. I used a quarter-turn connector to mate with the display (mating connection highlighted in green below) so that it could easily be removed to prevent theft. I suspended the connector from an arched two arm handlebar clamp with hex bolt fasteners.

FreeCAD-Camera-Handlebar

The mounting hardware was 3D printed with a high degree of infill to ensure a stable platform for testing. The hex bolts allowed for quick changes until the ideal viewing angle was achieved while in a neutral riding position. The display can be quickly attached or removed to the mounting system with a 90 degree turn of the connector.

mounting-hardware-design

The Final System and Lessons Learned

After securing the camera to the seat post I routed the cable under the top tube, installed the handlebar mount, and connected the display. After a few quick adjustments to get the proper alignment I took the rear-vision system for a test ride down my favorite country road.

Real-Time-Rear-Vision

The weeks of user testing and iterative design paid off moments later when I dodged a truck that decided to wander across the white line with its side-mirror and nearly strike me from behind. Unfortunately during that initial test run I didn’t have a microSD card installed in the prototype to capture any footage but it convinced me that a real-time rear-vision system empowers cyclists to be more than passive users of the road.

I learned a variety of lessons over the past few months while developing the prototype:

User Testing
It is critical to get user feedback ASAP. No matter how good a design looks or how well the electronics work if a device doesn’t solve the users problem it only compounds it further. The initial rear-view prototype experienced extreme glare and forced the rider to look down at the screen causing a distraction. Only user tests with different screen surface treatments and different handlebar positions resolved the issues.

Rapid Prototyping & Iterative Design
3D printers have revolutionized rapid prototyping and iterative design. In only a matter of hours I went from modified design concept in FreeCAD to a physical model mounted and tested on the road. The ability to make modifications and immediately test the new design permitted experimentation with a variety of configurations that might have otherwise been too costly to test.

Technical Limitations
There are always constraints that force trade-offs in a design. Initially I envisioned a slim and aerodynamic display enclosure but the physical constraints of the available electronics hardware dictated a specific thickness. Designing around technical limitations forced me to drill down into the requirements and prioritize the key features of the rear-view prototype.

Taking an initial idea and transforming it from concept to reality is a journey filled with new insights, challenges and opportunities. No matter what the idea, user testing and iterative design ensures real problems are solved and people enjoy the product experience.

SHARE ON:

Leave a Reply

Your email address will not be published. Required fields are marked *