Next-Gen Autopilot Puts A Robot At The Controls

August 25, 2023 by Bryan Cockfield 19 Comments

While the concept of automotive “autopilots” are still in their infancy, pretty much any aircraft larger than an ultralight will have some mechanism to at least hold a fixed course and altitude. Typically the autopilot system is built into the airplane’s controls, but this new system replaces the pilot themselves in a manner reminiscent of the movie Airplane.

The robot pilot, known as PIBOT, uses both AI and robotics technology to fly the airplane without altering the aircraft. Unlike a normal autopilot system, this one can be fed the aircraft’s manuals in natural language, understand them, and use that information to fly the airplane. That includes operating any of the aircraft’s cockpit controls, not just the control column and pedal assembly. Supposedly, the autopilot can handle everything from takeoff to landing, and operate capably during heavy turbulence.

The Korea Advanced Institute of Science and Technology (KAIST) research team that built the machine hopes that it will pave the way for more advanced autopilot systems, and although this one has only been tested in simulators so far it shows enormous promise, and even has certain capabilities that go far beyond human pilots’ abilities including the ability to remember a much wider variety of charts. The team also hopes to eventually migrate the technology to the land, especially military vehicles, although we’ve seen how challenging that can be already.

Go Fly A Kite

November 13, 2022 by Bryan Cockfield 17 Comments

Harvesting energy from the wind has been a commercially viable way of generating clean energy for around three decades now. Wind turbines are a reliable, proven technology but they do have some downsides, one of which is that since there’s more wind higher above the ground this usually means tall, expensive towers. There is a way around this problem, though, which is using kites to generate energy instead of a fixed turbine.

While kite generators aren’t a new idea, [Benjamin] has been working on this kite generator which has a number of improvements over existing kite generators. Like other kite generators, this one uses a tether to spin a generator which is located on the ground. But while this is similar to other kite systems, this prototype has a much simpler design and sweeps a much larger area while in flight. It also has an autopilot with multiple independent steering systems, which [Benjamin] says will allow it to stay in flight for months at a time provided there is enough wind. If there isn’t, it can land reliably, and launching it is relatively fast and simple as well.

While kites do have some obvious downsides compared to fixed turbines including a single point of failure at the tether and a large amount of cleared area to operate, they have plenty of advantages as well. They’re smaller, simpler, require no complicated yaw system, and can be easily maintained on the ground. In fact, it’s possible to build very simple kite generators out of nothing more than a hobby kite and some readily-available electrical components.

Continue reading “Go Fly A Kite” →

R2Home Is Ready To Bring Back Your High Altitude Payload

August 1, 2022 by Danie Conradie 23 Comments

With high-altitude ballooning, you are at the mercy of the winds, which can move your payload hundreds of kilometers and deposit it in some inaccessible spot. To solve this [Yohan Hadji] created R2Home, an autonomous parachute-based recovery system that can fly a payload to any specified landing site within its gliding range.

We first covered R2Home at the start of 2021, when he was still in the early experimental phases, but the project has matured massively since then. It just completed its longest and highest test flight. Descending autonomously from a release altitude of 3500 m, with an additional radiosonde payload, it landed within 5 m of the launch point.

R2Home electronics with it's insulated enclosure — R2Home electronics with its insulated enclosure

R2Home can fly using a variety of steerable canopies, even a DIY ram-air parachute, as demonstrated in an earlier version. [Yohan] is currently using a high-performance wing for RC paragliders.

A lot of effort went into developing a reliable parachute deployment system. The main canopy is packed carefully in a custom “Dbag”, which is attached to a drogue chute to stabilize the system during free-fall and deploy the main canopy at a preset altitude. This is done with a servo operated release mechanism, while steering is handled by a pair of modified winch servos intended for RC sailboats.

All the electronics are mounted on a stack of circular 3D printed brackets which fit in a tubular housing, bolted together with threaded rods. With the help of a design student [Yohan] also upgraded the simple tube housing to a lockable, foam-insulated design to help it handle temperatures at high altitudes.

The flight main flight computer is a Teensy 4.1 plugged into a custom PCB to connect all the navigation, communication, and flight systems. The custom Arduino-based autopilot takes inputs from a GPS receiver, and pilots the system to the desired drop zone, which it circles until touchdown.

The entire project is extremely well documented, and all the design files and code are open source and available on Github. Continue reading “R2Home Is Ready To Bring Back Your High Altitude Payload” →

Taking Reverse Engineering To The Skies: Cheap Drone Gets PX4 Autopilot

April 24, 2021 by Stephen Ogier 3 Comments

Sometimes bad software is all that is holding good hardware back. [Michael Melchior] wanted to scavenge some motors and propellers for another project, so he bought an inexpensive quadcopter intending to use it for parts. [Michael] was so surprised at the quality of the hardware contained in his $100 drone that he decided to reverse engineer his quadcopter and give the autopilot firmware a serious upgrade.

Upon stripping the drone down, [Michael] found that it came with a flight management unit based on the STM32F405RG, an Inertial Measurement Unit, magnetic compass, barometric pressure sensor, GPS, WiFi radio, camera with tilt, optical flow sensor, and ultrasonic distance sensor, plus batteries and charger! The flight management unit also had unpopulated headers for SWD, and—although the manufacturer’s firmware was protected from reading—write protection hadn’t been enabled, so [Michael] was free to flash his own firmware.

We highly recommend you take a look at [Michael]’s 10 part tour de force of reverse engineering which includes a man-in-the-middle attack with a Raspberry Pi to work out its WiFi communication, porting the open-source autopilot PX4 to the new airframe, and deciphering unknown serial protocols. There are even amusing shenanigans like putting batteries in the oven and freezer to help figure out which registers are used as temperature sensors. He achieves liftoff at the end, and we can’t wait to see what else he’s able to make it do in the future.

Of course, [Michael] is no stranger to hacking imported quadcopters, and if you’re interested in PX4 but want something quieter than a quadcopter, take a look at this autopilot-equipped glider.

GPS Guided Parachutes For High Altitude Balloons

January 7, 2021 by Danie Conradie 32 Comments

Most amateur high altitude balloon payloads descend back to earth with a simple non-steerable parachute and can land hundreds of kilometers from the launch site in inaccessible areas. [Yohan Hadji] experienced this first-hand during a balloon launch conducted by his high school, which inspired him to R2Home, a GPS-guided parachute recovery system.

A Teensy runs the show, and controls a pair of sail winch servos pulling the brake lines

[Yohan]’s first challenge was to create a steerable parachute that can deploy reliably, so he started doing tests with a borrowed scale model paragliding wing. He quickly learned that a canopy aspect ratio of below two was needed for reliable deployment, so he started sewing his own canopies. Steering a parachute involves pulling on a pair of brake lines, one for each side of the parachute. A control stroke of about 20 cm was required, and [Yohan] found that RC sailboat winch servos work perfectly for this application. The entire system is designed to fit in a 7×40 cm tube, and the parachute is deployed with the help of a small drogue chute and a servo-operated release mechanism.

[Yohan] is working on a custom flight controller, built around a Teensy 4.1, GPS receiver, and digital compass. A possible alternative is Ardupilot, which we’ve seen used on several autonomous drones, gliders, and rovers. While this system might not be possible to return to the launch point, it could certainly close the gap, and land safely in a designated area.

So far [Yohan] has done a series of test drops from a drone at low altitude to test deployment and steering, using an RC controller. The project is open source, and the mechanical design files and control code is up on GitHub. As with most 16-year-olds, [Yohan]’s resources are limited, so feel free to drop him some financial help on the R2Home GoFundMe page. See the videos after the break for a development montage and project presentation. Continue reading “GPS Guided Parachutes For High Altitude Balloons” →

Tesla Begins “Full Self Driving” Public Beta As Waymo And Cruise Go Unattended

October 23, 2020 by Lewin Day 131 Comments

Self-driving technology is a holy grail that promises to forever change the way we interact with cars. Thus far, there’s been plenty of hype and excitement, but full vehicles that remove the driver from the equation have remained far off. Tesla have long posited themselves as a market leader in this area, with their Autopilot technology allowing some limited autonomy on select highways. However, in a recent announcement, they have heralded the arrival of a new “Full Self Driving” ability for select beta testers in their early access program.

Taking Things Up A Notch

Telsa’s update notes highlight the new “Full Self-Driving” capabilities. Drivers are expected to pay continuous attention and be prepared to take over at any time, as the system “may do the wrong thing at the worst time.”

The new software update further extends the capabilities of Tesla vehicles to drive semi-autonomously. Despite the boastful “Full Self Driving” moniker, or FSD for short, it’s still classified as a Level 2 driving automation system, which relies on human intervention as a backup. This means that the driver must be paying attention and ready to take over in an instant, at all times. Users are instructed to keep their hands on the wheel at all times, but predictably, videos have already surfaced of users ignoring this measure.

The major difference between FSD and the previous Autopilot software is the ability to navigate city streets. Formerly, Tesla vehicles were only able to self-drive on highways, where the more regular flow of traffic is easier to handle. City streets introduce far greater complexity, with hazards like parked cars, pedestrians, bicycles, and complicated intersections. Unlike others in the field, who are investing heavily in LIDAR technology, Tesla’s system relies entirely on cameras and radar to navigate the world around it. Continue reading “Tesla Begins “Full Self Driving” Public Beta As Waymo And Cruise Go Unattended” →

Firmware Hints That Tesla’s Driver Camera Is Watching

October 13, 2020 by Tom Nardi 163 Comments

Currently, if you want to use the Autopilot or Self-Driving modes on a Tesla vehicle you need to keep your hands on the wheel at all times. That’s because, ultimately, the human driver is still the responsible party. Tesla is adamant about the fact that functions which allow the car to steer itself within a lane, avoid obstacles, and intelligently adjust its speed to match traffic all constitute a driver assistance system. If somebody figures out how to fool the wheel sensor and take a nap while their shiny new electric car is hurtling down the freeway, they want no part of it.

So it makes sense that the company’s official line regarding the driver-facing camera in the Model 3 and Model Y is that it’s there to record what the driver was doing in the seconds leading up to an impact. As explained in the release notes of the June 2020 firmware update, Tesla owners can opt-in to providing this data:

Help Tesla continue to develop safer vehicles by sharing camera data from your vehicle. This update will allow you to enable the built-in cabin camera above the rearview mirror. If enabled, Tesla will automatically capture images and a short video clip just prior to a collision or safety event to help engineers develop safety features and enhancements in the future.

But [green], who’s spent the last several years poking and prodding at the Tesla’s firmware and self-driving capabilities, recently found some compelling hints that there’s more to the story. As part of the vehicle’s image recognition system, which usually is tasked with picking up other vehicles or pedestrians, they found several interesting classes that don’t seem necessary given the official explanation of what the cabin camera is doing.

If all Tesla wanted was a few seconds of video uploaded to their offices each time one of their vehicles got into an accident, they wouldn’t need to be running image recognition configured to detect distracted drivers against it in real-time. While you could make the argument that this data would be useful to them, there would still be no reason to do it in the vehicle when it could be analyzed as part of the crash investigation. It seems far more likely that Tesla is laying the groundwork for a system that could give the vehicle another way of determining if the driver is paying attention.

Continue reading “Firmware Hints That Tesla’s Driver Camera Is Watching” →