Вы находитесь на странице: 1из 1

Modular Intelligent Driver Assistance System

By Derek Phillips

This project will design, build, program, and deploy
an intelligent automobile collision avoidance system that
leaves the driver in control of the vehicle at all times.
MIDAS will aid in reducing the staggering number of road
collision fatalities in conjunction with the UN Decade of
Action for Road Safety.
There are three aspects of the system: Sense the
environment, Determine the danger level, and Warn the
driver. The system itself is designed to fit on all models of
cars, a feature essential to success in countries where
people have a wide variety of cars, most of which are
unlikely to be modern.

Technical Implementation

Collaboration Methodology

The MIDAS combines the advances in sensor

hardware and intelligent detection to produce an easy-toinstall module that is attached to the roof of a vehicle,
regardless of model. The system takes in sensory
information from the environment, processes it, and then
may output a warning to the driver. The current design
includes a photovoltaic cell (PVC) to power the device.

A key component to this system is that it is developed

through extensive collaboration with established industry
companies, governmental and non-governmental
organizations. Many technology companies will be ecstatic to
get a chance to engage in global outreach, and as such most
of the components and required resources will be obtainable
at a reduced cost once an effective prototype and business
plan are presented.


Potential Trajectories

Connecting Rods


One of the most exciting aspects of this initiative is the

potential scalability and versatility of the system. Here are

Every year, nearly 1.3 million people die in road
accidents across the world, being the ninth leading cause
of death globally (World Health Organization). Additionally,
up to 50 million people suffer non-fatal injuries as a result
of car accidents, with many causing disabilities and
incurring severe losses to the productivity of society.
With the terrible results of car accidents in mind, and
the knowledge that ~90% of fatalities occur in low and
middle income countries, I set out to find a way to reduce
the number of collisions by leveraging newly developed
technology in the automotive industry. The focus of this
particular system is on pedestrian accidents, as they
account for a large portion of all deaths.
Figures from World Health Organization
Global Status Report on Road Safety, 2015

some of the simple-to-implement possibilities:

Add auditory warning messages in selected language
Sell a commercial version of the technology
Incorporate RADAR and LIDAR sensor systems
Remote updates to more advanced detection algorithms

In order to accurately perceive the environment 24
hours a day, the system will incorporate dual cameras,
each with both day and night vision. The cameras are
cheap relative to more advanced systems such as LIDAR
and RADAR, while still performing well in most situations.

Software will take the sensory input and output a
warning if the situation warrants a driver response. Most of
the computer vision algorithms already exist, and the
situation processing can be done by leveraging existing
algorithms in use by companies working on autonomous
vehicles. For example, Ford Motor Company already has
a pedestrian detection algorithm.

The MIDAS uses aspects of aircraft collision
avoidance systems in the design of the warning system. In
particular, if a dangerous situation is detected, the system
will indicate the recommended driver response.

Add braking and other vehicle control into MIDAS.

With approximately 300,000 road traffic deaths in 2015
in Africa alone, there is tremendous potential for the initiative
to save many lives. Even if only implemented in half of the
continent, and only targeting pedestrians, there is the
potential for over 100,000 lives to be saved. If the total
number of accidents is only decreased by 20%, this could
save an estimated 1% of gross national product in some of
the most heavily affected countries (WHO).

Plan of Action
2016: Develop detec0on and
danger evalua0on algorithm

2017-2020: Gain support from
BFI, Google, UN, etc.

2017-2020: Develop full

hardware/soJware prototype

2020: Gain good publicity through

product demonstra0ons

2021: Deployment and full
scale produc0on

2020-2021: Comprehensive
tes0ng and con0nual updates

Speed Up!

Slow Down!

Turn Left!

Turn Right!