Boeing sponsored research at MIT where Agha-mohammadi and pals have developed a new algorithm based on Partially Observable Markov Decision Processes (POMDP) that allows self monitoring of the vehicles health status in real time and alternative pathfinding, accessing recharging stations and landing zones if needed during mission execution. The paper will be presented the IEEE/RSJ International Conference on Intelligent Robots and Systems, in Chicago. Original article can be found here:
http://www.mit.edu/~aliagha/Web/pubpdfs/2014.Ali.Ure.ea.IROS_package.pdf
Comments
I should add - this is definitely interesting work when it comes to the idea of a ground vehicle supporting several UAVs to deliver various lightweight packages over short distances each time the ground vehicle stops in a new neighborhood. It'd be even more interesting in the ground vehicle ('base node' & 'recharge node') were moving. What if the ground vehicle didn't even bother to stop, but just drove slowly through town and sold ice cream while all the UAVs delivered packages? :)
Hopefully the author might read DIY Drones - how would the results change if the reactive planner was only slightly more conservative on when to return to base? 10% seems like a low reserve for a good comparison. It would be interesting to see how performance changes with a 20% to 50% threshold on RTB decisions.