In this project, we developed the worst idea of an automatic traffic light
assistanti which worked as a personal planning tool for people moving in a urban
environment. You can check the original business plan at https://venture-lab.org/venture/team_ideas/4586
The main change in business model is the shift from the original
consumer-oriented device to a city-planning oriented technology. From this
perspective, the customer becomes the administration of a middle-large city,
while the overall value proposition remains to improve the citizens life
(with a estimated reduction of time spent by motorists waiting for
lights to change by over 28%) quality, and the environmental impact of carbon emissions (statistics in
this domain demonstrate the possibility of a 6.5% reduction in CO2).
Key activities become the installation and monitoring of the devices, and the eventual fine tuning of the learning
algorithms for the artificial intelligence component (see below for tech
details). The cost structure is actually very low, consisting mainly in the
installation of the monitoring devices and initial training of the operators.
The recommended revenue stream is a periodic fee from city administrations,
which covers both computational and infrastructure upgrading costs.
A brief introduction to the idea behind the algorithm:
Actual traffic flow algorithms are developed around the approach of partial
differential equations in simulating the state of the city in a given
moment. The proposed approach would integrate pdes with retinal
algorithms to scan the effective amount of people and vehicles at a number
of adjacent crossings (ideally all the crossing of a high traffic city
area), and a neural network system trained to predict future states of the
system. This should allow to find light turning patterns which, while
assuring classical results in traffic control such as effective green waves,
constantly shifts to fit the actual needs of the users.
Post time: Jan-27-2017