Ironclad, one of many types of combat Uncrewed Ground Vehicle fielded by Ukrainian forces.<\/p>\n
Office of Mykhailo Federov<\/p>\n
Last month, Ukrainian officials announced plans to field 15,000 ground robots<\/a> \u2013 Uncrewed Ground Vehicles or UGVs — in 2025. That is a huge scaling up, and for obvious reasons.<\/p>\n \u201cIt\u2019s no secret that Ukraine is facing a severe shortage of personnel,\u201d Kateryna Bondar<\/a>, Fellow at Wadhwani AI Center, at the thinktank Center for Strategic and International Studies <\/a>(CSIS) told me. \u201cThis creates an urgent, existential need to substitute human roles with robotic systems.\u201d<\/p>\n Ukraine has scaled up drone production at pace, going from a few thousand in 2022 to 200,000 in 2024 to 2 million last year. But putting thousands of metal boots on the ground to replace humans is a far more challenging prospect, as Bondar explains.<\/p>\n Army Of Machines<\/p>\n The figure of 15,000 combat robots came from Hlib Kanevskyi, head of the Ukraine\u2019s Ministry of Defense quoted in Economichna Pravda. Kanevskyi says they signed contracts for $2.5 million of UGVs in the last half of 2024, in the first quarter of this year that surged to $150 million. This suggests a roughly hundred-fold increase.<\/p>\n There is certainly plenty of optimism about military robotics, and no shortage of Ukrainian UGV designs, with developers unveiling a new model every week or so. These are wheeled and tracked machines of various shapes and sizes, with roles from minelaying and mine clearing to logistics, casualty evacuation and direct combat either with explosive payloads or machine guns. Recent trials by the BRAVE1 defense technology incubator<\/a> involved 70 different types.<\/p>\n But there is much less sign of these machines at the front. Out of 50 types approved for military use, only 10 to 15 are in regular service according to . Much if this is due to the practicalities, which are more challenging than with shoebox-sized FPVs.<\/p>\n \u201cImagine bringing a 1,000-kilogram machine to the front line,\u201d says Bondar. \u201cIt\u2019s a logistics issue. If it runs on fuel you have to transport that as well, if it runs on batteries they will be big and heavy and require generators to recharge.\u201d<\/p>\n Then there is the cost. While small drones are literally expendable items costing a few hundred dollars apiece, UGVs are not so affordable.<\/p>\n \u201cEven a small one with a small payload starts at around $2,000-$3,000,\u201d says Bondar. \u201cThe more expensive ones are $10,000 and up. This is partly because they are much harder to manufacture — you cannot just make a body on a 3D printer.\u201d<\/p>\n At those prices, UGVs are only worthwhile if they can do things which aerial drones cannot. And given that drones are excellent for scouting and hitting targets many miles away with precision strikes, UGVs tend to be confined to more niche roles.<\/p>\n Machines Saving Lives<\/p>\n \u201cLogistics are an interesting use case,\u201d says Bondar. \u201cUGVs are now conducting way more missions and replacing people. Using a UGV removes a person from physical risk.\u201d<\/p>\n Zmiy – “Snake” \u2014 is a typical logistics UGV, a remote-controlled vehicle able to carry 1,100 pounds … More of cargo<\/p>\n Wikimedia Commons<\/p>\n The last few miles to the front line are within the strike range of Russian FPV drones, and trucks and vans are easier to attack than dug in positions.<\/p>\n \u201cGoing there and coming back is a super high-risk mission,\u201d says Bondar. \u201cA UGV which can carry 60 kilos can resupply a troop position for four or five days. That\u2019s very efficient.\u201d<\/p>\n Casualty evacuation going the other way is equally vital, though few are willing to trust a wounded soldier to a machine which may break down on the way and evacuation by UGV is seen as a last resort.<\/p>\n Ukrainian news sources described a UGV evacuation last month<\/a>, in which three wounded soldiers were transported more than 10 miles through an area covered by Russian mortar and artillery fire. The operation was a success, but it required more than 50 people to carry out, including the UGV operator, drone operators to track the UGVs progress, electronic warfare support and others.<\/p>\n This highlights another issue with UGVs: they may be uncrewed, but it takes a lot of people to operate them. And while driving up and down well-defined roads is difficult enough, it is even harder for combat robots which have to go off road. These typically require a driver plus as weapon operator to control the UGV, as well as a two-person drone team to provide aerial overwatch and guide the team on the ground. That\u2019s four people plus a load of hardware to replace one soldier.<\/p>\n A Russian drone operator shares his display with a colleague operating a UGV on the ground, giving … More the UGV operator an overhead view to drive his robot.<\/p>\n Russian State news media<\/p>\n An image of a Russian UGV team in action from a TV news report<\/a> illustrates this problem well. The UGV is in long grass, and the operator cannot see where he is going. The drone operator shares his display, so the UGV operator gets a bird\u2019s-eye view of his position \u2013 but the drone operator still needs the display to fly his own machine. Better software and machines which can direct themselves will greatly ease the workload.<\/p>\n \u201cThe future is all about autonomy and AI,\u201d says Bondar.<\/p>\n Smart software would allow the UGV to navigate for itself, using imagery from an autonomous drone flying overhead. The operator then becomes a mission commander, ordering the UGV where to go, and selecting and approving targets for the weapons. But achieving this is far more difficult than autonomous flight.<\/p>\n \u201cEven in civilian contexts, autonomous ground navigation remains unsolved; self-driving cars still struggle with edge cases on paved roads despite years of investment and defined traffic rules,\u201d says Bondar.<\/p>\n And of course, things are much more difficult in a military context.<\/p>\n \u201cThe system must be able to perceive its environment in real time, make context-aware decisions, avoid obstacles, and control the vehicle\u2019s complex mechanical systems\u2014all under conditions of GPS denial, degraded comms, and electronic warfare,\u201d says Bondar. \u201cWhile some promising prototypes exist in research labs, widespread battlefield deployment\u2014especially where a single human serves as a mission commander for multiple UGVs\u2014will likely take several more years to mature, even in high-urgency environments like Ukraine.\u201d<\/p>\n Robots On The Offensive<\/p>\n But while the software is still maturing, with plenty of human assistance the hardware can already take on carry out useful missions.<\/p>\n \u201cUkraine has already carried out their first robot-only assault,\u201d says Bondar. \u201cThe whole thing was completely remote-controlled.\u201d<\/p>\n This was a successful attack on Russian positions involving a mix of scout drones<\/a>, bombers, FPVs and dozens of UGVs on the ground carried out by the 13th National Guard Brigade at the end of last year.<\/p>\n Bondar notes that as soon as the UGVs had done their job, soldiers moved in to secure the position. The Ukrainians call this approach \u201cassault without assault<\/a>\u201d because the soldiers themselves are not involved in the action.<\/p>\n \u201cIt looks like Ukraine\u2019s main goal in this operation was to collect experience, then create guidelines and standards in term of tactics so this knowledge can be distributed,\u201d says Bondar. \u201cBut I\u2019m pretty sure we will see more of those.\u201d<\/p>\n Defensive operations are more straightforward. Ukraine has many robotic weapon systems and automated turrets. The simplest of them, are little more than a machine gun on a tripod with a video camera and a servo motor<\/a>. Operators using these can co-ordinate with aerial scouts and FPV operators to blunt Russian attacks without anyone in the trenches.<\/p>\n \u201cAt this point I think they could hold a defensive position fairly well,\u201d says Bondar.<\/p>\n Humanoid Warriors?<\/p>\n Looking forward, many companies are starting to produce humanoid robots which, in theory, could take on a role as soldiers. However, even tracked machines designed for the terrain have mobility issues.<\/p>\n \u201cOur robot got stuck in the mud while performing a task and we sent another robot with a winch to rescue it,\u201d runs a typical complaint cited in a Ukrainian report on UGV operations<\/a>. \u201cAs a result, the Russians destroyed the robot with the winch.\u201d<\/p>\n Mud, snow, sand and grass are frequently mentioned as major obstacles to UGVs.<\/p>\n ‘Robot dog’ being tested in Ukraine. These have not yet found a battlefield role <\/p>\n AFP via Getty Images<\/p>\n Legged robots, which are supposed to provide superior mobility over all sorts of terrain are not necessarily the answer. Ukraine has deployed quadrupeds \u2013 \u2018robot dogs\u2019<\/a> \u2013 but Bondar reports the results have been poor.<\/p>\n \u201cOperators complained that ploughed fields are a problem for quadrupeds,\u201d says Bondar. \u201cAnd they get stuck in the mud. And they can\u2019t hide. At this point they are a nice toy, with no application on the front line yet.\u201d<\/p>\n In principle, a humanoid robot would be far more mobile and could cross ground as easily as a human. Machines like Tesla\u2019s much-hyped Optimus<\/a> and Unitree\u2019s Humanoid G1<\/a> could be useful \u2013 when the AI that drives them is smart enough. Elon Musk suggests that Optimus will start from around $30k and will drop in price. But flashy demonstration videos are one thing, practicality \u2013 as with the robot dots \u2013 is another.<\/p>\n Machines like Tesla’s Core Technology Optimus humanoid robot could in theory offer a low-cost, … More mobile platform for military applications.<\/p>\n CFOTO\/Future Publishing via Getty Images<\/p>\n \u201cWhile in theory they offer human-like mobility and could eventually move, hide, and take cover like a soldier, their real utility will only become clear through practical battlefield testing,\u201d says Bondar.<\/p>\n Battery life of a couple of hours could be more than enough for a robotic unit to carry out an assault, and take a position which could then be occupied by human back-up. This is only a small tactical step from actions from assaults with tracked UGVs, with the difference that humanoids might advance faster and enter trenches and dugouts to clear them. In theory, at least.<\/p>\n \u201cUltimately, the affordability and effectiveness of humanoid robots will depend entirely on the type of missions they\u2019re tasked with and how reliably they can be reused,\u201d says Bondar. \u201cUntil those use cases are tested and proven, their true value remains speculative.\u201d<\/p>\n Ukraine\u2019s 15,000 robots will be working alongside over 800,000 humans, and for the time being the machines will be in a supporting role rather than taking over combat.<\/p>\n \u201cUGVs are likely to carve out an essential role in performing the most dangerous support tasks, where they can reduce risk to human soldiers without being lost in unsustainable numbers,\u201d says Bondar.<\/p>\n