If you see what looks like an ambulance with a set of flashing lights weaving through Tokyo traffic, it might not be rushing to save a human. GMO Internet Group is unleashing what it calls Japan's first "humanoid ambulance" on the streets of Tokyo this month. A maintenance van designed to make house calls when humanoid robots break down, the ambulance carries diagnostic equipment, spare parts, repair tools and engineers to wherever a robot has stopped working. If a robot can't be fixed at the roadside, there's a spare humanoid on board ready to be subbed in and take over on the spot. Meanwhile, the broken unit is ferried back to GMO's repair base at its Humanoid Lab in Shibuya, Tokyo.

Follow this section to personalize your feed and get instant alerts. Follow Go to your personalized feed WHY FOLLOW? Smart Alerts: Get notified about major news as it happens. Follow this tag to personalize your feed and get instant alerts. Follow Go to your personalized feed WHY FOLLOW?

Stella Juva, a solar vehicle built by Eindhoven University students, functions as a mobile clinic that powers its own medical equipment using sunlight in remote areas.

'I ran because I knew I would die': Russian drones target medics in Ukraine Inna Lytvynenko always puts on her bright orange body armour when she's out on an emergency call. Last June, the paramedic was sent out in Kherson to treat a woman injured in a Russian drone attack, when an FPV drone smashed into her ambulance. Seconds later, she heard a buzzing sound in the sky. Another drone was flying straight towards her, fast. I grabbed my trauma bag and started running, she recalls.

Simulation plays a key role in scaling robot learning and validating policies, but constructing simulations remains a labor-intensive process. This paper introduces ReGen, a generative simulation framework that automates simulation design via inverse design. Given a robot's behavior -- such as a motion trajectory or an objective function -- and its textual description, ReGen infers plausible scenarios and environments that could have caused the behavior. ReGen leverages large language models to synthesize scenarios by expanding a directed graph that encodes cause-and-effect relationships, relevant entities, and their properties. This structured graph is then translated into a symbolic program, which configures and executes a robot simulation environment. Our framework supports (i) augmenting simulations based on ego-agent behaviors, (ii) controllable, counterfactual scenario generation, (iii) reasoning about agent cognition and mental states, and (iv) reasoning with distinct sensing modalities, such as braking due to faulty GPS signals. We demonstrate ReGen in autonomous driving and robot manipulation tasks, generating more diverse, complex simulated environments compared to existing simulations with high success rates, and enabling controllable generation for corner cases. This approach enhances the validation of robot policies and supports data or simulation augmentation, advancing scalable robot learning for improved generalization and robustness. We provide code and example videos at: https://regen-sim.github.io/

While simulators exist for vehicular IoT nodes communicating with the Cloud through Edge nodes in a fully-simulated osmotic architecture, they often lack support for dynamic agent planning and optimisation to minimise vehicular battery consumption while ensuring fair communication times. Addressing these challenges requires extending current simulator architectures with AI algorithms for both traffic prediction and dynamic agent planning. This paper presents an extension of SimulatorOrchestrator (SO) to meet these requirements. Preliminary results over a realistic urban dataset show that utilising vehicular planning algorithms can lead to improved battery and QoS performance compared with traditional shortest path algorithms. The additional inclusion of desirability areas enabled more ambulances to be routed to their target destinations while utilising less energy to do so, compared to traditional and weighted algorithms without desirability considerations.

In Figure 9 we provide bias and MSE analysis of different algorithms on the domains that exhibit passive non-stationarity.

Kyiv, Ukraine – As luck would have it, emergency doctor Elina Dovzhenko was far enough from her vehicle when a Russian drone struck it, breaking the windshield and splattering pieces of shrapnel around. It was getting dark on July 9 in the bombed-out, nearly-abandoned city of Kupiansk which sits less than 5km (3 miles) from the front line in the northeastern Ukrainian region of Kharkiv – and just 40km (25 miles) west of the Russian border. But there was definitely enough light left for the Russian drone operator on the front line's opposite side to see that Dovzhenko's vehicle was a white ambulance with red stripes parked near a shelling-damaged hospital where she and her colleagues were. "We heard the drone move, it swirled and swirled around [the building], then we heard the blast," Dovzhenko, 29, told Al Jazeera. She and her colleagues were shocked and angry – but not surprised.