Keeping Badgers At Bay With Tensorflow

Human-animal conflict is always a contentious issue, and finding ways to prevent damage without causing harm to the animals often requires creative solutions. [James Milward] needed a humane method to stop badgers and foxes from uprooting his garden, leading him to create the Furbinator 3000, a system that combines computer vision with audio deterrents..

[James] initially tried using scent repellents (which were ignored) and blocking access to his garden (resulting in more digging), but found some success with commercial ultrasonic audio repellent devices. However, these had to be manually turned off during the day to avoid annoying activation of the PIR motion sensors by [James] and his family, and the integrated solar panels couldn’t keep up with the load.

This presented a good opportunity to try his hand at practical machine vision. He already had a substantial number of sample images from the Ring cameras in his garden, which he turned into a functional TensorFlow Lite model with about 2.5 hours of training. He linked it with event-activated RTSP streams from his Ring cameras using the ring-mqtt library. To minimize false positives on stationary objects, he incorporated a motion filter into the processing pipeline. When it identifies a fox or badger with reasonable accuracy, it generates an MQTT event.

[James] modified the ultrasonic devices so they would react to these events using an ESP8266-based WeMos D1 Mini Pro development board and added an external 5 V power supply for sustained operation. All development was performed in a Docker container which simplified deployment on a Raspberry Pi 4.

After implementing the system, [James] woke up to the satisfying sight of his garden remaining untouched overnight, a victory that even earned him some coverage by the BBC.

Thanks for the tip [Laurent]!

Autonomous Wheelchair Lets Jetson Do The Driving

Compared to their manual counterparts, electric wheelchairs are far less demanding to operate, as the user doesn’t need to have upper body strength normally required to turn the wheels. But even a motorized wheelchair needs some kind of input from the user to control it, which still may pose a considerable challenge depending on the individual’s specific abilities.

Hoping to improve on the situation, [Kabilan KB] has developed a self-driving electric wheelchair that can navigate around obstacles by feeding the output of an Intel RealSense Depth Camera and LiDAR module into a Jetson Nano Developer Kit running OpenCV. To control the actual motors, the Jetson is connected to an Arduino which in turn is wired into a common L298N motor driver board.

As [Kabilan] explains on the NVIDIA Blog, he specifically chose off-the-shelf components and the most affordable electric wheelchair he could find to bring the total cost of the project as low as possible. An undergraduate from the Karunya Institute of Technology and Sciences in Coimbatore, India, he notes that this sort of assistive technology is usually only available to more affluent patients. With his cost-saving measures, he hopes to address that imbalance.

While automatic obstacle avoidance would already be a big help for many users, [Kabilan] imagines improved software taking things a step further. For example, a user could simply press a button to indicate which room of the house they want to move to, and the chair could drive itself there automatically. With increasingly powerful single-board computers and the state of open source self-driving technology steadily improving, it’s not hard to imagine a future where this kind of technology is commonplace.

A hot tub with a smartphone in front showing real-time sensor data

ESP32 Keeps Track Of Hot Tub’s Vital Signs

Like swimming pools, hot tubs need regular monitoring to ensure their water stays clean and clear. An average person might take a water quality reading once or twice a week using test strips, but such a low sampling rate obviously won’t do for a hacker. [Stephen Carey] has therefore built a hot tub monitor that checks the water quality every minute and reports it on a neat mobile dashboard.

[Stephen]’s system uses commercially available sensors that track pH levels and Oxidation-Reduction Potential (ORP), both basic measurements that indicate water quality. A second set of sensors keeps track of the temperature of the water and the outside air, which should help in finding insulation failures and keeping energy use under control.

A set of graphs showing a hot tub's pH and ORP over time, with a significant spike in both near the beginningAn ESP32 reads the sensors and sends out the data through WiFi. [Stephen] programmed the ESP32 in MicroPython, using an MQTT driver to connect it to Home Assistant. By looking at the graphs generated, you can tell when someone entered the tub from a step change in pH and ORP. It’s even possible to generate alerts when any of the values drift outside their acceptable range – we can already imagine an alarm going off when someone enters without having showered first.

The system also has a calibration mode to check the sensors against a well-defined buffer solution. As with many chemical sensors, the pH and ORP probes gradually lose their active material and need to be replaced after about a year. Good ones aren’t cheap, but [Stephen] has found pretty decent low-cost alternatives on AliExpress that should be fine for a home setup.

If you also want your tub or pool to be actively managed, you’ll need a more complex system, perhaps even one that can also dispense chemicals. If your hot tub is heated by a wood fire, however, all you need is a way to alert the person tending the fire.

Open Source Tracker Keeps An Eye On Furry Friends

Most of the time, you’ll know where your cats are — asleep on the bed about 23.5 hours a day and eating or pooping the rest of the time. But some cats are more active than others, so there’s commercial options for those who want to keep tabs on their pet. Unfortunately, [Sahas Chitlange] didn’t like any of them, so he designed and built his own open source version: FindMyCat.io.

The system is in two parts: a module that fits onto a cat collar, and a home station that, well, stays at home. It offers a variety of tracking modes. In home mode, the home station signals the collar every 10 seconds, which stays in a deep sleep most of the time. If the collar doesn’t get a signal from the home station, it switches to ping mode, where it will wait for a signal from the FindMyCat over the LTE-M connection and report its location.

Finally, the app can set the collar to Lost Kitteh mode, where the collar will send a location to the app every seven minutes or thirty seconds. The collar also supports a direction-finding feature, using the ultra wideband (UWB) feature of recent Apple iPhones to point you in the direction and distance of the tracked cat.

The collar is built around a Nordic Semiconductor NRF-9160, a System in a Package (SiP) that does most of the heavy lifting as it includes GPS, an LTE-M modem, and an ARM processor. One interesting feature here: [Sahas] doesn’t make his antennas on the PCB, but instead uses an Ignion NN03-310, an off-the-shelf antenna that is already qualified for LTE-M use. That means this system can be connected to almost any LTE-M network without getting yelled at for using unqualified hardware and making the local cell towers explode.

The collar also includes a DWM3001CDK ultrawideband (UWB) module used for the locator feature. The accompanying app uses this and Apple’s UWB support to show the user which direction the cat is in, and how far away it is. The app isn’t in the Apple App Store yet, so you’ll need to sign up for an Apple Developer account to use it. We’d love to hear from anyone who takes it for a test drive with their own pet.

Continue reading “Open Source Tracker Keeps An Eye On Furry Friends”

Toy Bulldozer Becomes Epic Terrifying Lawnmower

Regular lawnmowers are a perfectly fine way to mow your lawn, but they can be a bit boring. They’re also not always the best at tackling thick brush and bushes. [rctestflight] has a solution to both of those problems, in the form of a plant-munching bulldozer.

The concept is simple — it starts with a hefty miniature RC bulldozer. Weighing in at 27 kilograms (60 pounds), the beast has actual functioning hydraulics to control the blade and plow. It struggles somewhat with traction, particularly in muddier conditions, and can’t really dig much, but it nonetheless looks the business.

As cool as it was, [rctestflight] decided to employ it for some real yard work by outfitting it with a mowing rig. The ‘dozer was outfitted with a pair of sawblades, run by twin brushless motors for plenty of grunt. That gave the bulldozer the ability to mow through not just lawn, but even thick blackberry bushes and two-foot high weeds.

It’s not great at steering, but it’s able to destroy thick brush with reckless abandon. Fundamentally, it looks like a very fun way to mow an overgrown yard.

Continue reading “Toy Bulldozer Becomes Epic Terrifying Lawnmower”

Grannophone Helps You Stay In Touch

Whether it’s distance, pandemics, or both that separate you from your elderly loved ones, what’s the best idea for communicating with them so they don’t suffer from loneliness on top of issues like dementia? We’d say it’s probably something like [Stefan Baur]’s Grannophone.

Back in late 2020, a Twitter user named [Nitek] asked the Internet what could be done in the way of a grandma-friendly video-conferencing solution, provided Grandma has a TV and a broadband internet connection. At first, [Stefan] was like, just get her an old iPad and FaceTime with her. But the question got him thinking. And prototyping.

Grannophones are essentially Linux machines with a video-capable SIP client connected over a VPN for privacy reasons. In simple mode, picking up the handset of one Grannophone will call the other, but more complicated configurations are possible. We particularly like that replacing the handset automatically obscures the camera. That’s a nice touch.

At this point, the Grannophone is a work in progress. The idea is that they be extremely easy to build at the kitchen table, like on the order of disposable Swedish furniture. If you can contribute to the project, please do. Be sure to check out the demonstration video after the break.

On the other hand, if Granny is 1337, you could always video-conference in terminal.

Continue reading “Grannophone Helps You Stay In Touch”

An automatic laser turret playing with a cat.

Entertain Your Cats Automatically With LazerPaw

Most of us would agree that kittens are very cute, but require lots of attention in return. What would you do if you adopted three abandoned cats but didn’t have all day to play with them? [Hoani Bryson] solved his problem by building LazerPaw — an autonomous, safe way to let your cats chase lasers.

Having recently tinkered with computer vision in the form of OpenCV, [Hoani] decided he would make a laser turret for his cats to play with. An infrared camera, used so that the LazerPaw works in the dark, is mounted to the laser and the Raspberry Pi. These electronics are then mounted on a servo-based pan/tilt module, which is in turn mounted with two smartphone clamps to the ceiling. That way, when the cats chase the laser, they will be looking away from the beam source. Additionally, if the device is aiming directly at a cat, the laser is turned off. Finally, [Hoani] added some NeoPixels with an Arduino-based controller for extra hacker vibes.

The LazerPaw’s software takes in a 30 FPS stream from a webcam, scales it down for performance, and applies a threshold filter to it. When a black pixel, which is assumed to be a cat, is detected, it “pushes” the camera away from it depending on how close to the laser it is. The effect of this is that every time a cat catches up to the laser, it moves away again. The processed images are also sent to an interactive website for remote cat playtime. Finally, there is also a physical start button so you don’t need WiFi to use it.

Is your cat more of a sunbather than a deadly murder beast? Maybe it’ll like this cat chair that follows the sun.

Continue reading “Entertain Your Cats Automatically With LazerPaw”