Weekly Robotics Newsletter Archive [5/13]

  • Weekly Robotics #32

    0) Weekly Robotics Survey.
    Google
    INFO: If you have 5 minutes to spare we would greatly appreciate your input in the above survey. Your input is valuable to us and will help us create a better experience for you. Thanks for being with us!

    1) rosshow.
    GitHub
    INFO: rosshow is a tool for visualizing ROS (Robot Operating System) topics in a terminal. Among other things it supports point clouds, images, laser scans, imu and nav sat messages.

    2) NASA-funded Inflatable Robots are Safe in Space, and in Homes.
    YouTube
    INFO: Brigham Young University engineers are working on programmable inflatable robots. The advantages of inflatable robots mentioned in the video are: robots taking less space (and weight) while not in use, safety, ability to absorb impacts.

    3) Airbus Begins Shore-to-ship Port Drone Deliveries.
    AIN online
    INFO: Airbus has launched shore to ship trials in Singapore. During trials Skyways, a delivery drone capable of lifting up to 4 kg (8.8 pounds) delivering cargo to vessels as far as 1.6 nautical miles (3 km) from the coast, was used. During trials the drone delivered 1.5 kg (3.3 pounds) of 3D printed parts to a ship 0.8 nautical miles away from the shore.

    4) Optimal Actuator Design.
    biomimetics.mit.edu
    INFO: In this post biomimetics lab (the lab that created MIT Cheetah robot presented in Weekly Robotics #28) presents their requirements for an optimal actuator. The team ended up designing their own actuator based on Emoteq HT-5001 motor and an 5.8:1 planetary gear.

    5) GITAI Partners With JAXA to Send Telepresence Robots to Space.
    IEEE
    INFO: GITAI and JAXA are looking into creating telepresence robots for space activities. The robots would be working on space stations on low earth orbit, while being teleoperated from Earth. In the article you can find a video with the GITAI robot being teleoperated through a VR system.

    6) Handle Robot Reimagined for Logistics.
    YouTube
    INFO: In this YouTube video Boston Dynamics showcases it’s mobile robot designed for logistics applications. The robot is designed to handle boxes as heavy as 15 kg (33 pounds).

    7) Publication of the week - TossingBot: Learning to Throw Arbitrary Objects with Residual Physics (2019).
    arxiv.org
    INFO: This paper by researchers affiliated with Princeton University, Google, Columbia University and MIT showcases TossingBot, an UR5 robot capable of grasping arbitrary objects from unstructured bin and throwing them into target bins outside of manipulator’s workspace at 500+ mean picks an hour. The neural network developed in this project consists of perception, grasping and throwing modules. The system uses an RGB-D camera for creating a height map (later used by grasping and throwing modules) and an overhead RGB-D camera that is used to detect changes in the landing zone before and after the throw. The system achieves 84.7% accuracy in real world conditions for the objects seen during training. You can find more information on the project, including videos on the project website.

    Careers

    1) Rovco (Bristol, UK) - Robotics Engineer.
    INFO: Rovco is a fast-growing subsea robotics company focused in the renewables and offshore decommissioning sectors. We are advancing the state of the art in subsea robotics by developing the technology for live 3D vision and machine learning based data analysis.

    2) Neobotix (Heilbronn, Germany) - Software developer – ROS.
    INFO: Neobotix develops and manufactures mobile autonomous robots for professional service robotics research. Our portfolio ranges from intelligent autonomous vehicles to mobile manipulators to tailor-made robots forhighly specialised applications.

    Announcements

    0) Introducing Announcement section.
    Weekly Robotics
    INFO: We are introducing an announcement section dedicated to robotics related announcements such as workshops, events, grants and more! If you would like your organization to be featured in this section then please get in touch with us.

    1) ROS-I EU Spring ’19 Workshop.
    rosin-project.eu
    INFO: ROS-Industrial EU Tech Workshop will take place from May 6th to May 7th 2019 in Stuttgart, Germany. The workshop will focus on ROS 2, specificaly DDS/RTPS implementation, microROS and poriting of ur_modern_driver to ROS 2. The workshop is free for members of any ROS-Industrial Consortium or 500 Euro for individuals from other organizations.

  • Weekly Robotics #31

    1) PathFinding.js.
    qiao.github.io/
    INFO: PathFindings.js is an open source project providing path finding algorithm for web based games. The page above provides an interactive visualization of some of the algorithms provided by the library.

    2) Robot Hand is Soft and Strong.
    MIT
    INFO: MIT researchers created an origami shaped, vacuum-powered gripper that can lift objects that are up to 100 times heavier than its weight. You can see the gripper in action on YouTube.

    3) Autonomous Snowplow Makes Debut on Winnipeg Runway.
    CTV News
    INFO: Northstar Robotics developed an autonomous snow plower that is meant to be used at an airport. According to this post the machine is powered by ROS.

    4) Space Gateway.
    ESA
    INFO: In 2020s the partners of the International Space Station will launch a Space Gateway - a staging post much higher than ISS that could be used as an intermediate point for astronauts before going on the missions deeper into space. The gateway will weight 40 tonnes and the astronauts will be able to occupy it for up to 90 days at a time.

    5) Jetson Nano.
    NVIDIA
    INFO: NVIDIA announced Jetson Nano, a 99$ single board computer for AI and Robotics. The board has dimensions of 70 x 45 mm and according to the above website the board will require up to 10 watts of power. Jetson Nano will be available from distributors worldwide from June 2019. If you get one and want to install ROS Melodic on it then you might like this tutorial by Stereo Labs.

    6) Robotic Dreams, Robotic Realities: Why Is It So Hard to Build Profitable Robot Companies?.
    IEEE
    INFO: This article from IEEE talks about how difficult it is to be profitable in robotics and how we should build realistic expectations for our products and avoid overselling.

    7) Publication of the week - Snapbot: a Reconfigurable Legged Robot (2017).
    Disney Research
    INFO: This paper introduces a modular legged robot design that can use up to 6 legs. The legs come at 2 DOF (roll-pitch or yaw-pitch) or 3 DOF (roll-yaw-pitch) variants, creating 700 possible robot configurations. The robot has snap-on connectors with 4 neodymium magnets and pogo-pin electrical connectors. The connector design combined with providing each servo with an ID allows for configuration discovery (performed every 100ms). The robot supports three types of motion: rowing, crawling and working, all depending on the leg configuration. You can see the video presenting the robot on YouTube.

    Careers

    1) Latent Logic (Oxford, UK) - Robotics Engineer.
    INFO: Latent Logic’s mission is to use state-of-the-art A.I. to learn complex human behaviours from real demonstrations, such as video data, so that we can design autonomous systems which can work collaboratively alongside humans.

  • Weekly Robotics #30

    1) Why Anti-Poaching Drones Haven’t Worked Well.
    DIYDrones
    INFO: This post on DIY Drones by Chris Anderson highlights a discussion from ArduPilot Discourse on the topic of using drones for anti poaching efforts. The discussions are interesting both from the engineering and environmental perspective. As user Graham_Dyer says in one of his post: “From my experience I believe, that even with the best intentions and best technology, drones will never have more than a negligible effect on poaching. To put it in context, the Canine teams in the KNP’s best haul was 18 poachers apprehended in one week. Our drone team saw less than 8 poachers in TWO years of night flying, with not one apprehension. So if the drone is not CATCHING an average of 2-5 poachers a WEEK then it’s of no use. And note that poachers get around deterrents easily”.

    2) 3 New Chips to Help Robots Find Their Way Around.
    IEEE
    INFO: This article from IEEE highlights advances in systems-on-chip (SoC) that can perform robot navigation and path planning and even visual SLAM (Simultaneous Localization and Mapping). The systems described in the article were developed by engineers and researchers from Intel, University of Minnesota and University of Michigan.

    3) The MIT Super Mini Cheetah.
    MIT
    INFO: In issue #28 we’ve presented MIT Mini Cheetah. The Super Mini Cheetah is another project from MIT Biomimetics Lab. Compared to the larger version the Super Mini is a low cost build that has simpler kinematics and in our opinion resembles a Cheetah a lot more than it’s bigger brother (it might have to do with the robot’s head).

    4) Supervised Racing.
    markku.ai
    INFO: This blog post from January describes Donkeycar project (presented in Weekly Robotics #2) AI. We think that the post gives a nice big picture of how Donkeycar works. The team behind Markku recently open sourced a set of 3D designs that can be used for attaching the required hardware to the R/C car.

    5) A Rare Look at the Meltdown Inside Fukushima Daiichi Nuclear Power Plant.
    CNET
    INFO: This article from CNET shows how robots are used in cleaning up Fukushima nuclear power plant. Most of the work done there by robot is exploratory - giving technicians input on the conditions inside of the reactor. According to the article engineers are working on custom robotic systems that can be used in the high radiation level zones. The article contains many photos of the Fukushima plant, including photos taken at Naraha Center for Remote Control Technology, where Japan Atomic Energy Agency setup facilities for engineers to test robots designed to work at the nuclear site.

    6) OpenMV - Low Cost Machine Vision Module.
    openmv.io
    INFO: “The OpenMV project is about creating low-cost, extensible, Python powered, machine vision modules and aims at becoming the “Arduino of Machine Vision“. Our goal is to bring machine vision algorithms closer to makers and hobbyists. We’ve done the difficult and time-consuming algorithm work for you leaving more time for your creativity”. The project is open sourced under MIT licence. On GitHub you will find both the hardware schematics and software.

    7) Publication of the week - Differential-Drive In-Pipe Robot for Moving Inside Urban Gas Pipelines (2005).
    ent.mrt.ac.lk
    INFO: This paper by Se-gon Roh and Hyouk Ryeol Choi introduces MRINSPECT IV, a differential-drive robot meant to be used for 10 cm (4 in) pipes inspection. The length of the robot is 150 mm (6 in), the minimum radial size it can achieve is 85 mm (3.3 in) and the maximum is 109 mm (4.3 in). The weight of the robot is 0.7 kg (1.5 lbs) and it can travel at a speed of up to 0.15 m/s. The paper provides interesting insights on using a differential drive in complex geometric workspaces. In the paper the researchers focus on geometric analysis of the behaviour and wheel velocity control in branches and elbows.

    Careers

    1) Maritime Robotics (Asker/Oslo, Norway) - Electrical Engineer.
    INFO: Maritime Robotics is a leading provider of innovative unmanned solutions for maritime operations in harsh environments. Our systems operate unmanned in the air and on the surface.

    2) Canonical (home based, US) - Robotics Engineer.
    INFO: Ubuntu is the operating system of choice for robotics. We at Canonical (the company behind Ubuntu) want to make it even better, and we’re looking for talented developers to join the effort.

  • Weekly Robotics #29

    1) Google Summer of Code 2019.
    Google
    INFO: “Google Summer of Code is a global program focused on bringing more student developers into open source software development”. The Robotics and Space related projects students can contribute to this year are: ArduPilot, JdeRobot, RoboComp, Open Source Robotics Foundation, JSK Robotics Laboratory, Open Roberta, OpenCV, RTEMS Project, Libre Space Foundation, GNSS-SDR, BRL-CAD. If we’ve missed any projects please send us an e-mail and we will update the list.

    2) Drones At Sea: The Highs & Lows of a Parley SnotBot Expedition.
    DroneLife
    INFO: This article from DroneLife talks about Ocean Alliance usage of drones for whale monitoring in a SnotBot expedition. Even though the team faced numerous challenges they managed to achieve the expedition objective of capturing over 50 biological samples.

    3) Optical Gyro Dwarfed by a Grain of Rice.
    Elektor
    INFO: Caltech engineers developed an optical gyroscope the size of a grain rice. From the article we’ve learned that the optical gyroscopes make us of Sagnac effect (Wikipedia, Simple simulation on YouTube and are usually the size of a golf ball.

    4) Bare Metal STM32 Programming and a Quadcopters Awakening.
    timakro.de
    INFO: In this blog post Tim Schumacher describes his experience in programming an STM32 chip onboard Crazypony mini, a 100$ mini quadrotor. Even though the author only shows how to control a single LED this article can be a good entry point for anyone looking into programming embedded autopilots.

    5) Vehicle Path Planning.
    Razor Code
    INFO: This article on vehicle path planning by David Olsen talks a lot about heuristics of path planning and optimization. It also showcases an interesting vehicle model and approaches to path optimization.

    6) I Flew a Helicopter, and then the Helicopter Flew Me.
    The Verge
    INFO: Piloting a helicopter after 45 minutes of training on simulator? It is now possible with SARA (Sikorsky Autonomy Research Unit). The system is fit with a fly-by-wire system allowing the pilot to control the aircraft using two inceptors (a lever and a joystick). The system comes with an HMI (Human Machine Interface) in form of a screen with widgets, allowing the pilot to plan flight parameters that the aircraft will execute automatically.

    7) Publication of the week - Autonomous Drifting using Simulation-Aided Reinforcement Learning (2016).
    WeeklyRobotics
    INFO: This paper by Mark Cutler and Jonathan P. How introduces a framework that combines simple and complex continuous state-action simulators with a real-world robot to efficiently find good control policies, while minimizing the number of samples needed from the physical robot. While many Reinforcement Learning Researchers use expert demonstration to initialize the models, the authors of this paper chose an approach in which they generate initial policies from simple models. On the car Researchers track velocities in x and y direction (the information is provided by a motion capture system), turn rate and wheel speed. By applying the algorithm shown in this paper combined with prior simulated information the R/C car was able to drift as it can be seen in this YouTube video. You can download additional material to the paper here.

    Careers

    1) Saildrone (Alameda, CA, US) - Various Positions.
    INFO: Saildrone’s mission is to create the highest resolution ocean data set in the world and use it to make global processes such as weather forecasting, carbon cycling, global fishing, and climate change more predictable, visible, and actionable. We’ve linked to their Unmanned Surface Vehicle and Arctic mission in Weekly Robotics #28.

    2) Remy Robotics (Barcelona, Spain / Moscow, Russia) - Various Positions.
    INFO: Remy Robotics is using robotic technology to take the routine and inefficiency out of cooking, creating a ‘robot chef’ that can help cook any cuisine or any dish in the world.

    3) Samsung (Mountain View, CA, US) - Robotics Open Source Engineer.
    INFO: Open Source Group (OSG) under Samsung Research is the corporate Open Source Program office for all open source activities in Samsung.