Ethics and Solutions for Autonomous Vehicles

Amy Guo, Hetav Pandya, Jessie Lin

Type: Hybrid Solution

Main Concerns

Research

• 95% of police agree that drowsy driving is a serious problem. [1]

​

• Commuters waste an average of 54 hours a year stalled in traffic, study says. [2]

​

• Human Reliability on AV [3]
• Vehicle more automated → Driver more distracted
• Long adjustment times
• AVs are prone to be abused

Our solution focuses on addressing some of the major concerns that happen during the current transitional period towards fully automated vehicles…

​

Improvements are needed to support both the AI and improve human-computer interaction.

Motivation

• Autonomous vehicles are a expensive to buy
• Non-autonomous vehicles are a majority on the roads
• Based on the case study: Ethical issues in the autonomous industry

Giving non-autonomous vehicles the autonomous features that enhance safety without the need for huge capital investments on buying a new autonomous vehicle and give ethical control to humans.

Technical Solution (s)

Technical Solution (s)

• Focuses on adding features without altering the vehicle.
• Usage of AI to create a safer driving environment.

Dynamic Traffic Lights

Road signal detector

Drowsy Driver Detection

Automated Traffic Signals

• Used to identify amount of traffic and adjust traffic light duration.

​

• Controlling the flux of cars in an intersection.

Video demo of base code

Automated Traffic Signals

• Used to identify amount of traffic and adjust traffic light duration.

​

• Simulation developed using PyGame

Video simulation of use case

Road sign detection

• Camera usage to accurately detect road and traffic signs.

​

• Recognition software is currently trained on four different road signs.

Performance of our AI on the Test set

Driver Drowsiness Detection

• Camera usage to identify whether the driver is drowsy or awake
• Provided to users as a mobile app.

Video demo of base code

Policy Solution

Implementing the technology and addressing legal concerns

Policy Implementation

• Municipally Implemented Automated Traffic Signals
• Must go through rigorous testing
• Introduced to the public in stages
• Ensure privacy is maintained

​

• Traffic Signal Detection + Drowsiness Detection
• Not publicly provided, but encouraged for public safety
• Use in police/law enforcement

​

Data Privacy

• Consent
• Option to disable the software

​

• Transparency
• All data stored can visualized in real time

​

• Integrity and Access
• Temporary storage which works on a frame-by-frame basis.
• Completely offline functionality

​

PRINCIPLES IDENTIFIED IN

LAYING DOWN HARMONISED RULES ON ARTIFICIAL INTELLIGENCE (ARTIFICIAL INTELLIGENCE ACT) AND AMENDING CERTAIN UNION LEGISLATIVE ACTS

Based on European Union’s Artificial Intelligence Act

Sources

[1] https://tirf.ca/wp-content/uploads/2017/02/2009_Facts_Fatigue_Driving_Ontario_Police_Guidebook.pdf

​

[2] https://www.cnn.com/2019/08/22/us/traffic-commute-gridlock-transportation-study-trnd/index.html

​

[3]https://www.sciencedirect.com/science/article/pii/S0040162520309197?casa_token=-KGvcrMK9RkAAAAA:Lw3yOtQx5VHD85kLlnpr_OgfwChkdZux9jdhHQ0ykt0xibvm7F2PnyH_2lkuMoPw4hFpwFlWO5Gl

​

[4] https://www.sciencedirect.com/science/article/abs/pii/S002243751730381X

​

[5] https://www.idtechex.com/en/research-report/autonomous-cars-robotaxis-and-sensors-2022-2042/832

Providing Accountabilit

• Transparency of AV decision-making
• Helps law enforcement in the case of an accident
• Requirement for entering Canadian market

​

• Professionals in law and ethics must assist in design of AVs.

​