top of page


The conversational user interface in the driverless police vehicle 

Project Overview

The value of the conversational user interface (CUI) has been recognized in the driverless vehicles. While many CUI design focus on improving the driving experience, my team was interested in exploring how CUI can enhance the professional working experience, like the police officer's daily work. In this project page, you would see how we identified the problem, ideated and iterated the design solution, and how we did the experience prototyping.

​This is a two-week course project. My role in this three-people team project: 

  • Conducted the user research;

  • Designed the key features;

  • Created the experience prototyping; 

  • Designed the UI interface.

  • Competitive Analysis

  • User Interviews

  • User Journey Maps

  • Brainstorming 

  • Experience Prototying 

  • UI Design 

Frame the Problem

  • Competitive Analysis 

The Driverless Car

To learn more about what is currently available in terms of driverless cars, we cross compared three products that are currently being compared as driverless. The first product is the Uber self-driving taxis, which are currently being tested in Pittsburgh, the second is the newer cars that Tesla sells, and the third product is Waymo, an autonomous car that Google is producing.


Self-Driving Uber


  • Designed for Taxi Purposes

  • Fully Autonomous

  • No Voice Commands

  • Has Steering Wheel




  • Luxury Car for Consumers

  • Semi-Autonomous

  • Basic CUI tasks

  • Large Touch-Screen Interface


Waymo (Google)


  • Designed for Transportation

  • Fully Autonomous

  • No Voice Commands

  • Forwarding Mobility

  • No Steering Wheel

The CUIs

We studied the interactive forms of the CUIs and the mainstream Smart Home devices those CUI embedded in app/platform.

Interactive Forms



  • Voice

  • Text

  • Voice

  • Text

  • Sound

  • Light

Smart Home Devices


  • Listens for keywords

  • Built-in cloud

  • Crowd-powered: Free-to-use

  • API & Developer community

Google Home

  • More context-aware

  • Touch interface

  • Ability to connect with other smart devices


  • Designed for specific tasks

  • Textual input

  • Usually embedded in app/platform

  • User Interviews

We concluded that we would need more field knowledge to be able to get a better understanding about police officers' routines and duties, as well as the current technology they are using inside the police vehicle. We interviewed a Carnegie Mellon University campus police officer: Officer Sergeant Nellon. 

Interview Findings:

Driving & Patrolling

  • Focus on driving safely, limited attention

  • Usually just 1 driver, unless with a new officer in training

Current Tech in Police Vehicle

  • Mobile Data Terminal (MDT)

    • Computing device (computer) inside vehicle
    • Touch interface
    • Connected to database

        Pain point:

It takes a long time for authentication to run plate in database Need to change password every 30 days

  • Camera tied to MDT

  • LPR: license plate recognition

  • Radar:

    Only state police are allowed to use radar tech, not municipal

  • “Stop ambush”:

    2 sensors on the front and back of the vehicle
    When someone approaches the vehicle while the of cer is writing report, etc., a light flashes indicating in which way the other person is approaching

  • Machine to issue citation/warning:

    Swipe card to automatically ll in and print citation

Routines & Duties (process of a traffic stop)

  • See traffic violation

  • Turn on the emergency light

  • Traffic stop - position the vehicle behind the target within a certain distance

  • Dispatch to other police so database has a record, also for backup

  • Ask the driver for his/her information

  • Walk back to the police vehicle and run the information in database

  • Issue a citation/warning to the driver

  • Turn off the emergency light

  • Paperwork: write a report

Officer Nellon was activating the MDT inside the vehicle

Officer Nellon was demonstrating how he used the MDT when he was on duty

The interface of the MDT

Key insights we gained from the interview: 

  • Police officers are not able to pay attention to their surroundings while they are interacting with the MDT screen, which may pose safety hazards.

  • Police officers do not like interacting with the MDT screen because it usually takes a long time for them to log in and perform any duties.

  • Police patrol officers’ duties include not only patrolling and conducting traffic stops, but also writing report and other mundane tasks, which freeze their attention from their surrounding.

Based on the competitive analysis and the user interview , we defined the problem statement as: 

frame the problem

Police pays lots of attention on driving while they are supposed to pay primary attention to surveying their surroundings and identifying dangerous situations.

Solution Exploration 

  • User Journey Map

  • Brainstorming

Working towards our final scenarios

solution exploration

From our ideation process, our solution to the problem is: 

To free the police from intense multitasking (e.g. looking at the screens to check the information and observing the surroundings at the same time), we want to design a CUI that can increase safety and efficiency for both police officers and people around them. Also, help police make better decisions as their cognitive load is significantly decreased by using CUI.


The core value of our design is to support the police officer to make more appropriate decisions based on the more accurate and efficient communication with CUI.

Solution Validation 

  • Experience Prototyping & UI Design 

To illustrate our design idea, we created three scenarios and performed a live acting in class. The police refer to the CUI as PI (Police Interface) and allows them to get in contact with the CUI in two syllables. We hope PI can help anticipate and evaluate potential hazards, and provide efficient and accurate information support to police officers.

Although the digital prototype of the CUI is not required as the final deliverable, I designed the conversational user interface to visualize the features described in the user cases. 

When officers perform routine traffic stops, they are often in a lot of danger because cars can hit them and it is hard for them to scan their surroundings. This scenario shows how PI can increase the safety by helping the police officer focus on the suspect.


Officer Lucy pulled over a red car on the 376 highway at 3 am. She got the driver’s license and entered the data into the system.




PI, I scanned the driver’s ID, please check if he has any existing crime record.




Yes, officer Lucy. Checking now.

(Lucy’s eyes not on the car, focus on something else)



Alert! ID 34827, the name in this record is recognized an extremely dangerous suspect. He just made bail for having an illegal weapon and has had several DUIs this year. Please monitor him closely and remain vigilant, I will call back up.



Does he has other crimes on his record?



Checking now. Yes, he has a record of being violent. Make sure you stay in a safe place,

backup is on the way.



Got it, Thanks! I am waiting for backup.

Scenario 1
Freeing Attention
to Increase Safety

Visualize the image and the text data

Highlight the

key information

Multimedia Alert

According to our interview, It is common for officers to break out a routine to perform certain tasks in their daily work. This scenario demonstrates how PI helps the police officer to quickly respond to the circumstances.


Officer Sara is using a driverless vehicle to complete a routine patrol around Pittsburgh to look for any suspicious or unsafe activities.




Officer Sara, I have received alert from your supervisor, we are looking for a theft suspect. She was last seen on Morewood and Fifth.



Great, how can we get there as fast as possible?



I have found the fastest route to Morewood and Fifth. The vehicle will have to go the wrong

way on Morewood avenue. Please confirm.



Confirm. Let’s make sure we alert those around us!



I am activating the sirens now to alert pedestrians and the vehicle’s surrounding us. We are looking for a white female wearing a blue sweater with a green stolen purse.



Got it, thanks.

Scenario 2
Freeing Attention
to Increase Efficiency

Make the suspect information handy

Highlight the

map information 

Multimedia Alert

The error recovery is very important ability of CUI. This scenario presents how the error could happen and how it could be recovered through the interaction.


Officer Lucy receives a call from her supervisor that she will be dispatched to Morewood and Center avenue for patrolling.




PI, take me to Morewood and Center.



Sure. Heading to Morewood and Center Street right now.



Wait? Stop, PI. This is the wrong way.



Okay, would you like to go somewhere else?



I want to go to Morewood and Center avenue.



Oh, sorry about that. Heading to Morewood and Center avenue now.





Scenario 3
Error Recovery

Visualize the changed destination and the route

Make quick

error recovery 

bottom of page