Energy autonomy is one of the main challenges in robot exploration. The Marsu fleet, a marsupial robot society, is composed of two kinds of robots, the Marsu-bots (explorers) and the Mother-bot (a mobile charging station). The Marsu-bots are able to connect to each other and recharge batteries if necessary. A protocol will be implemented on the Marsu fleet so when a Marsu-bot runs out of battery another Marsu-bot in the fleet will move to its location and recharge its battery. Such a protocol has two main tasks: first, determine how good a robot may be at solving this task, and second, create a distributed algorithm that allows the Marsu-bots to autonomously select the best one. In this paper it will be shown that leader electionalgorithms can be used to implement the negotiation. It will also be shown that, while it is impossible to accurately predict the performance of a rescue operation based on distance and battery alone, it is possible to find a cost function that allows the selection of a robot able to perform the rescue operation without further assistance. Rules for defining the specific function to be used according to the scenario in which the robots are present will also be given.