Hi, I’m Diarmid. I’m a senior software engineer and I am currently working on simulating scenarios showing how Fetch.AI brings value to people.
The goal is to understand in detail how Fetch.AI Autonomous Economic Agents will behave in the real world so we can build our technology to support this behaviour. In addition, we are able to produce videos of the simulations to help the wider community understand the scope of our ambitions.
In the simulation below we are looking at the problem of charging electric cars on long journeys. People usually leave their electric cars to charge overnight, but if they need to recharge en route, drivers can face a long wait as it can take an hour or more to fully recharge their vehicle.
The scenario on the left with the orange cars represents common human behaviour. The electric cars in this part of the demo travel until they run low on power. At this point they check their Sat Nav system and look for charging stations that are within range. They pick the one which is the shortest detour from their main route and go there.
The difficulty for the human drivers is that when they arrive they may find that lots of other drivers have had the same idea and a long queue has already formed. A queue at an electric charging station is much more problematic than one at a traditional petrol station because you may have to wait for hours just to get to the front of the queue.
The scenario on the right-hand side of the video with the blue cars represents a Fetch.AI-enhanced scenario. The cars contain a Fetch.AI agent which is monitoring the car battery’s power level and is able to communicate with other agents on the Fetch.AI network. The charging stations also have a Fetch.AI agent monitoring the current waiting time for cars at the station.
The car’s agent monitors the battery and when it runs low on charge it searches for charging stations that are within range (so far, the same as the human behaviour). The car’s agent then seeks information about the current wait time at nearby stations from each charging station’s agent. The car’s agent then calculates the total detour time by adding up the time taken to get to the station, the wait time and the time to drive from the station to the destination. It can then choose the station offering the shortest detour overall (after factoring in the predicted time that would be spent waiting at a charging station).
As the car is on its way to the charging station, it continually monitors the wait times of the local stations and if the “shortest detour” station stops being the shortest, it begins the process again.
This inter-agent communication drastically reduces wait times at the charging stations. It also allows charging stations off the main roads to attract more customers.
There are many simplifications in this simulation. The most obvious is that, in reality, if a human arrived at a charging station and saw a 20-hour wait, they would go looking for an alternative station. However, we expect the kind of benefits we see on our simulation to carry across to the real world.
In our simulation, cars that didn’t queue at all completed the journey in about 8 hours and 30 minutes. The cars without the Fetch.AI agents took an average of 13 hours and 30 minutes. Cars with Fetch.AI agents took approximately 9 hours and 30 minutes — a 30% reduction in the time it took cars that weren’t using Fetch.AI’s technology.
This use case is also highly relevant when demonstrating the potential of machine verification. In a recent article, Fetch.AI’s CTO Toby Simpson outlined the importance of being able to prove your identity in today’s digital world. Machine verification is already prevalent in society and it will be adopted increasingly widely for everyday actions in the future. To take one example using electric vehicles, charging points owned by particular car manufacturers will want to know they can trust the vehicle they are being plugged into.
Fetch.AI recognises the importance of collaboration to deliver the IoT economy of the future. We are currently working with Outlier Ventures and Sovrin to develop ANVIL. The software allows individuals to seamlessly prove their identity to trusted representatives. This is an exciting step forward and we’ll be showcasing the technology and how it will function on the Fetch.AI network during a live webinar demonstration tomorrow (28 March). To find out more, please sign up here.