Robots operating alongside humans often encounter unfamiliar environments that make autonomous task completion challenging. Though improving models and increasing dataset size can enhance a robot's performance in unseen environments, dataset generation and model refinement may be impractical in every unfamiliar environment. Approaches that utilize human demonstration through manual operation can aid in generalizing to these unfamiliar environments, but often require significant human effort and expertise to achieve satisfactory task performance. To address these challenges, we propose leveraging part-time human interaction for redirection of robots during failed task execution. We train a lightweight help policy that allows robots to learn when to proceed autonomously or request human assistance at times of uncertainty. By incorporating part-time human intervention, robots recover quickly from their mistakes. Our best performing policy yields a 20 percent increase in path-length weighted success with only a 21 percent human interaction ratio. This approach provides a practical means for robots to interact and learn from humans in real-world settings, facilitating effective task completion without the need for significant human intervention.

Tesla's are becoming less like bats every day. The automaker has announced that it has started to remove the ultrasonic systems its current vehicles use for their electronic driver aid systems, including Autopilot and Full Self-Driving, which will transition to a solely camera-based system. Tesla began removing radar from its vehicles last year and the Model 3 and Model Y will be the first to lose the ultrasonic sensors, followed by the Model S and Model X next year. A "vision-based occupancy network" will draw information from eight cameras installed on the vehicles. Tesla's Autopilot and Full Self-Driving systems originally relied on an array of radars, ultrasonic sensors and cameras.

Interest in semi-autonomous systems (SAS) is growing rapidly as a paradigm to deploy autonomous systems in domains that require occasional reliance on humans. This paradigm allows service robots or autonomous vehicles to operate at varying levels of autonomy and offer safety in situations that require human judgment. We propose an introspective model of autonomy that is learned and updated online through experience and dictates the extent to which the agent can act autonomously in any given situation. We define a competence-aware system (CAS) that explicitly models its own proficiency at different levels of autonomy and the available human feedback. A CAS learns to adjust its level of autonomy based on experience to maximize overall efficiency, factoring in the cost of human assistance. We analyze the convergence properties of CAS and provide experimental results for robot delivery and autonomous driving domains that demonstrate the benefits of the approach.

There has been a lot of discussion about autonomous vehicles on the land and in the air, but what about on the sea? While the world got the first glimpse of a fully autonomous ferry thanks to the collaboration between Rolls-Royce and Finferries, the state-owned ferry operator of Finland, there's still quite a bit of work to be done before we can expect the world's waterways to be overtaken with autonomous vessels. Even though we might be years or even decades away from the majority of vessels becoming autonomous, there are certainly artificial intelligence algorithms at work today. A fully autonomous ship would be considered a vessel that can operate on its own without a crew. Remote ships are those that are operated by a human from shore, and an automated ship runs software that manages its movements. As the technology matures, more types of ships will likely transition from being manned to having some autonomous capabilities.

Getting to the future of transportation requires a good roadmap, and not in the metaphorical sense. Self-driving cars rely on robust knowledge of the world. So do less aspirational technologies: telling drivers how to get across town, tracking burgeoning networks of electric car chargers, geofencing cars with semi-autonomous systems to roads where they can stay safe. And every cartographer faces the same challenge: How to maintain the accuracy of a sprawling map in a world that can change from one moment to the next? Today, Mapbox, which supports map features for companies like Snap, Tinder, and the Weather Channel, announced its bid to do keep tabs on a world that doesn't stop moving.

Were you aware that 3,285 people were killed today, in a war that has been going on for decades, and that each year at least 20 million people have been injured in that same war? Well, it is not surprising, as this particular war is not well publicized, and rarely makes the news. This is the war that takes place between frail-bodied people -- often women and children -- and that unforgiving machine of destruction, the motor car. If that many people were killed and injured in a war anywhere in the world, it would definitely make headline news. As it is, the carnage is swept up, the bodies buried, and everyone tries not to notice.

Were you aware that 3,285 people were killed today, in a war that has been going on for decades, and that each year at least 20 million people have been injured in that same war? Well, it is not surprising, as this particular war is not well publicised, and rarely makes the news. This is the war that takes place between frail bodied people, often women, and children, and that unforgiving machine of destruction, the motor car. If that many people were killed, and injured, in a war anywhere in the world, it would definitely make headline news. As it is, the carnage is swept up, the bodies buried, and everyone tries not to notice.

On Monday, a Tesla Model S slammed into the back of a stopped firetruck on the 405 freeway in Los Angeles County. The driver apparently told the fire department the car was in Autopilot mode at the time. The crash highlighted the shortcomings of the increasingly common semi-autonomous systems that let cars drive themselves in limited conditions. This surprisingly non-deadly debacle also raises a technical question: How is it possible that one of the most advanced driving systems on the planet doesn't see a freaking fire truck, dead ahead? Tesla didn't confirm the car was running Autopilot at the time of the crash, but its manual does warn that the system is ill-equipped to handle this exact sort of situation: "Traffic-Aware Cruise Control cannot detect all objects and may not brake/decelerate for stationary vehicles, especially in situations when you are driving over 50 mph (80 km/h) and a vehicle you are following moves out of your driving path and a stationary vehicle or object is in front of you instead."

The great promise of autonomous vehicles, aside from saving you from the tyranny of commuting, is their ability to save lives by replacing stupid humans with intelligent computers. But these cars, at least in the short-term, could make driving riskier because people don't yet understand the technology or just how it works. British auto insurance companies call this "autonomous ambiguity," and it is not an abstract issue. Automakers like Audi, Cadillac, Mercedes-Benz, Tesla, and Volvo already or will very soon offer vehicles that do some of the driving for you. In a new white paper, the Association of British Insurers argues that drivers don't understand the limitations of these semi-autonomous systems, and believe their car is more capable than it really is. "This risk of autonomous ambiguity could result in a short term increase in crashes," said Peter Shaw, CEO of Thatcham Research which collaborated on the report.

If you're wondering why you don't have a self-driving car yet, you should know that you're part of the problem. When it comes to semi-autonomous systems--cars that do some of the driving, but leave the tricky stuff to humans--the biggest hurdle isn't the technical challenge of making a car safely drive itself. It's ensuring that you're alert and ready to grab the wheel if a sensor craps out, someone cuts you off, or any number of other sudden and random things happens. Engineers call this the handoff problem, and it's so tricky that companies like Ford and Google's Waymo have given up on partial autonomy because they figure it's easier to go straight to full autonomy and cut you out of the equation entirely. Tesla has done it pretty well with its Autopilot system, for example.