To that end, costs must be kept low. The defence department plans to pay less than $800,000 a pop for Gremlins, though that would be for an order of 1,000 of them. So, if a couple were shot down in an operation, "that's okay, it really doesn't matter," says Mr Wierzbanowski. In military jargon, he describes the things as "attritable". Dynetics' working assumption is that each Gremlin will fly a maximum of about 20 missions. this means they can be made from less durable, and therefore cheaper materials and components.
为此必须保持低成本。美国国防部计划为一架小精灵花费不超过80万美元，不过那是订购1000架的价格。所以，如果一次行动中有几架小精灵被击落，“没关系，真的不要紧”，维尔兹巴诺夫斯基说。用军事术语来说，他称小精灵是“可消耗的”。Dynetics初步假设每架小精灵最多能执行大约20次飞行任务。这意味着它们的制造材料和部件可以不那么耐用，价格也就更低。
The aerial aircraft-carrier of choice for the Gremlin project is a modified C-130 cargo plane, which could carry up to four of the drones in bomb racks slung under its wings. That would make for a small squadron, but numbers could be bolstered by further drones dropped from accompanying fighters or bombers. Deploying the drones is therefore fairly easy. The tricky part is fishing them out of the air when they return from a mission. For this, Dynetics has designed a special recovery system that fits above a C-130's cargo ramp.
小精灵项目首选的空中航母是经改造的C-130运输机，它机翼下方悬挂的炸弹架最多可以搭载四架无人机。这将形成一个小型中队，但随着同行的战斗机或轰炸机投放更多的无人机，总体数量就会增加。因此无人机的投放还算容易。棘手的部分是当它们完成任务返回时把它们从空中打捞回来。为此，Dynetics设计了一种安装在C-130运输机货舱门上方的特殊回收系统。
When a Gremlin flies back to the mothership, the cargo ramp opens and the recovery system lowers a boom out of it. This boom releases a pod on a ten-metre-long tether, and that pod clamps onto a short engagement arm which pops out of the top of the Gremlin itself. A successful capture shuts off the Gremlin's engine. A winch then hoists the drone on board. This arrangement should be able to pull eight Gremlins an hour out of the air.
当小精灵飞回母舰时，货舱门打开，从回收系统里放下一根吊杆。吊杆释放出一个用十米长的线缆连接的吊舱，吊舱抓住从小精灵顶部弹出的一个短的啮合臂。捕获成功后，小精灵的发动机就会关闭，再由绞盘将它拉上飞机。这种方式应该能在一小时内从空中拉回八架小精灵。
That, at least, is the idea. So far, though Gremlins have managed to come within centimetres of successful capture, such capture has not been achieved. But Dynetics hopes some software tweaks will deal with this by the summer, when operational testing of the Gremlins system by the air force is supposed to begin.
至少想法是这样的。到目前为止，小精灵尚未被成功捕获过——还差那么几厘米。但Dynetics希望一些软件上的改进可以在今年夏天之前解决这个问题，届时美国空军应该会启动对小精灵系统的实操测试。
This testing will include flights with various sorts of payloads. But the most important thing to be tested will be how well Gremlins are able to co-operate—for instance, by swapping tasks as circumstances evolve. For this, the air force is developing software that emerged from a different DARPA programme, called Collaborative Operations in Denied Environment. It will not be able to orchestrate a fully autonomous "swarm", at least for now. The goal, rather, is to give individual drones enough autonomy for a single human operator to be able to oversee a cluster of them.
这次测试将包括携带各种有效载荷的飞行。但最重要的是测试小精灵们的协作能力，比如随环境变化而交换任务。为此美国空军正在开发一款软件，它来自DARPA的另一个项目，名为“拒止环境中的协同作战”。该软件并不能组织一个完全自主的“机群”，至少目前不能。确切地说，它的目标是让单架无人机具备足够的自主行动力，这样一名人类操作员就能够监管一群无人机。
