Battlefield robots were once thought of as the stuff of science fiction. But the Ukraine war has put paid to that view, with uncrewed ground vehicles (UGVs) – or ground robots – taking on an increasingly visible role.
Author
- Jacob Parakilas
Research Leader, Defence, Security, and Justice Group, RAND Europe
In September 2026, Mykhailo Fedorov, the former defence minister of Ukraine – who is generally credited for modernising the country’s use of military drones – called for an army of robots on the frontline. He also outlined an ambitious challenge : achieve the first “confirmed battlefield result using a humanoid robot” within six months.
Ukraine is not alone in exploring the potential of humanoid machines. In the wake of the widely publicised World Humanoid Robot Games, a kind of Olympics for bipedal robots, held recently in Beijing, Reuters reported that the Chinese military is working towards developing humanoid robots for combat.
The US military has also been exploring the potential use of humanoid robots on the battlefield. According to the BBC , San Francisco-based Foundation Robotics has picked up US$24 million (£18 million) in research contracts to pilot its humanoid technology with the US military. Two of its Phantom Mk1 robots have reportedly also been undergoing tests in Ukraine .
In the age of artificial intelligence, it would be foolish to completely dismiss the prospect of rapid technological advancement. But while the concept of humanoid robots for the battlefield grabs attention, there are a number of substantial obstacles to this becoming a near term reality.
In general, though, battlefield ground robotics is developing rapidly. In Ukraine, Uncrewed Ground Vehicles (UGVs) which move on wheels or tracks have taken on an increasing range of tasks. These include the evacuation of casualties, combat engineering and even, in limited circumstances, offensive action . Drones are also used to supply troops with rations, weapons and ammunition.
Ukraine recently commemorated their contribution to the battlefield with an Independence Day robot parade through the centre of Kyiv.
In principle, a bipedal robot resembling a human would have the flexibility of an infantry soldier -, able to run, jump, crawl, climb, or potentially even swim across obstacles. This means that it could be used for a range of tasks in distinct environments.
In short, a bipedal robot offers flexibility and adaptability as its key characteristics.
But the flexibility and adaptability of the human form – driven by millions of years of evolution towards making us into a hardy, survivable all-purpose species rather than a specialist predator or herbivore – has not yet been matched by mechanical engineering.
And while the Ukrainian battlefield is not the sole indicator of what the future of war looks like, it does remind us of crucial limiting factors on technological progress. Cost matters hugely, systems matter more than individual weapons, and the most advanced and transformative technology is not necessarily the most photogenic.
Bipedal drawbacks
There are two basic problems with legged robots that even the most advanced designs are not close to overcoming in the near future.
The first is that legs are complicated. Wheels and tank tracks are both mechanically simpler than legs, which gives them an inherent resistance to damage or breakdown. It’s also worth noting that a wheeled or tracked vehicle can travel lower to the ground, which reduces its visibility and exposure to enemy fire.
Both, as we have seen in Ukraine, are matters of survivability . The relative simplicity of tracks or wheels also makes them cheaper to mass produce – and the military that chooses a small quantity of legged robots is likely to be rolled over by one that chooses plentiful, if less technically impressive, wheeled or tracked units.
These are all factors that can be mitigated by engineering. But that engineering adds development cost, mechanical complexity and, very likely, unit cost. These are not insurmountable obstacles, but they are significant ones.
The other problem, which will be familiar to anyone who owns or has driven an electric vehicle, is the limited range. After all, relatively few soldiering tasks can be accomplished while plugged in. There are some where it’s feasible to imagine range being a minor issue. For example, a humanoid robot could presumably be assigned to guard a static facility where it would walk a prepared path and be able to recharge regularly.
But that task could also be accomplished by proven, simpler, cheaper methods, such as a combination of sensors, obstacles, and remote-controlled weapons. This undermines the case for deploying bipedal robot sentries.
In a jam
Outside such controlled environments, the open battlefield is not notable for its proliferation of reliable, high-powered recharging stations or the provision of frequent breaks so that combatants can revive themselves.
A further problem – not unique to ground robots, but exacerbated by the challenges inherent in traversing the ground – is maintaining control. Despite recent advancements in autonomy, for the foreseeable future uncrewed ground vehicles such as those being used on Ukraine’s frontline, will need live links to their human controllers . Basically, that means a choice between a radio link or a physical fibre optic cable link.
The general balance here is that fibre optics cannot be jammed and provide extremely high data transfer rates and high fidelity. But they have limited range and are susceptible to being severed by terrain, obstacles, or blasts and shrapnel. Relative to flying drones, the range limitation issue is less severe and the danger of physical severance more severe for ground robots.
By contrast, radio control is susceptible to jamming, electronic interference and disruption over distance and in inclement weather. They can also be blocked by obstacles. But a radio link cannot be physically severed. Repeaters are devices that receive signals and re-transmit them, so that two-way radio signals can cover longer distances. These provide high degrees of range and operational flexibility for remote systems.
Increasingly, remote systems are being designed with both types of control available. This, for example, allows the radio system to serve as a backup in the event that the fibre optic cable is severed. But doing so adds cost and complexity.
These factors may seem mundane, but when taken together, they should add a degree of realism to the discussion around a technological area which has long since captured the popular imagination. After all, how many cinema tickets would have gone unsold if Robocop was taking orders via a trailing fibre optic cable and stopped the film every half hour because his battery had run low?
The technology behind humanoid robots is undoubtedly impressive – and in time it may prove transformative. But the engineering, physical and operational obstacles to humanoid soldier robots remain stubbornly in place.
