CNTRL/REACT: Wireless joystick with force feedback

A modular drone or robot motion controller. Reaction wheel based system providing haptic feedback. Can be used with only one hand.

Public Chat
Similar projects worth following
While the idea of a gyroscope-enabled control "stick" is nothing new, and has even been released as a commercial product, our implementation tries to address a big problem with such solutions, by bringing back the physical feedback and tactile "detents" that our brains use to intuitively gauge how much input is given at any point in time, without distracting the pilot/operator.
The R/C industry "standard issue" controller with two gimbals has remained mostly unchanged since its wide adoption in the 80's. This project challenges the status quo and aims to bring a solution that can be used with a single hand and in a more intuitive way.

The existing gyroscope-enabled controllers don't include a feedback system. They lack any physical feedback, as the one created by the spring tension in classical RC gimbals and wired joysticks. One main issue is the inability of the pilot to perfectly maintain the zero position, causing unwanted, erroneous inputs.

However, one hand wireless joysticks do have a lot of advantages over classical "gimbal" controllers. Besides the obvious accessibility features, we have a more ergonomic grip and more intuitive control, since the movements of the hand translate directly into movements on the craft/device. 

In our initial testing, we have found that this device can flatten the learning curve for new drone pilots, since rewiring your brain and hand movements is not necessary. From my observations, new pilots tend to tilt the head or body in the directions in which they intend to move, but then slowly translate the movements into stick inputs. This step almost entirely skipped with our device.

Since the case and trigger assembly is based on the stellar work from the  Tarantola NG by ikletti,  some modularity and future upgrade paths will be available. Customisation is also an important factor in human interface design, and we want to provide the end user with the capability to customise and modify their own parts, or at least choose from a range of parts that are screwed together.

feedback system demo TRMD.mp4

A small demo of the haptic feedback system. Initially we did not expect it to be able to move the hand by this much, and was aiming just for a feeling or sensation. However the system proved to be more than capable, and sometimes even shaking loose. Threadlocker reccomended!! Future versions might get away with smaller or less powerful motors.

MPEG-4 Video - 45.88 MB - 08/23/2021 at 14:46


  • 1 × ESP32 dev board. Brains of the operation. Possible future replacement is the Arduino nano 33 BLE.
  • 1 × MPU6050 IMU board Used to determine the angle of the device. primary input method.
  • 1 × "Tarantola NG" 3D printed case. Initially intended as an R/C car controller, based on the Taranis QX7 drone controller, we adapt it to our needs. -
  • 1 × Taranis QX7 replacement gimbal. Hall effect or salavage. Used for a smooth and precise trigger assembly. 3D printed files available from the Tarantola project.
  • 1 × M3 screws. Used for attaching different case parts together and for mounting the motors.

View all 8 components

  • Methods of control

    [ E C C 0 ]08/23/2021 at 19:41 0 comments

    There are three main ways in which you can control a drone or an R/C airplane. Here we’ll compare all the advantages and disadvantages of these input methods. Especially concerning portability, since this is a necessary feature for RC aircraft/drone control. As we will see, these input technologies are quite similar, but there are a few differences that we need to account for.

    The WIRED JOYSTICK is a hand-controlled device in which any deflection of the stick from the centre point is considered input. It is composed of two parts: the stable base, which does not move, and the control stick, which moves in relation to the base. The angle of the deflection is measured by a pair of potentiometers, or, in certain cases, a pair of hall-effect sensors reading the rotational position of a magnet that is affixed on the control stick, translated into an analog value (further digitised by an ADC). Since we have two perpendicular axes of motion, the TWO output values will represent deflection in the X axis, and in the Y axis, respectively. This defines a plane of possible inputs, usually represented as a 2D plane, and a point representing the current position of the stick. Important characteristics of this device are as follows: A big input window, self-centering, comfortable/ergonomic, NOT portable.

    A GIMBAL is a finger-controlled, miniaturised version of a joystick, with the same functioning principle. Even though it uses the same technology, potentiometers or hall-effect sensors (even the same part numbers as their bigger counterpart in some cases), some mechanical differences appear. Because of the limited space in which the gimbal has to fit, and because it is supposed to be controlled with fingers, some compromises had to be made. One of the biggest compromises is the reduction in the size of the input window. R/C gimbals can be up to 5 times smaller than a standard joystick. This creates the issue of accuracy. Since the input window is smaller, a smaller human error will translate into a bigger error in the data, resulting in sending unwanted input to the craft.

    The important characteristics of this device are: small input window, self-centering, less comfortable/ergonomic, PORTABLE.

    A third option exists, in the form of WIRELESS JOYSTICKS, which, even if very similar in shape and function to their wired counterpart, use a totally different technology for recording data. The wireless (or portable) joystick can not have a stable base, as it would be hard to store, carry, and use while in the field, since it needs a flat and stable surface as a reference. This would limit the use cases to places with outside tables, or would force the user to sit on the ground, severely limiting mobility and possibilities. This is why wireless joysticks do not have a base at all, incorporating an Inertial Measurement Unit (IMU) including an electronic gyroscope and accelerometer in the control stick itself. The data collected from this IMU is usually then sent through a radio channel to a PC for processing, or directly to the craft to be used as control data.

    Since the IMU does not directly determine the angular deflection of the joystick, some processing is necessary. This processing and filtering can reduce the accuracy of the measurement, but if done correctly, this should not pose a big problem. The notable characteristics of this device are: big input window, comfortable/ ergonomic, portable, but NOT SELF-CENTERING.

    Even though on paper, this looks like a better option than the RC gimbal, because of the heightened possible precision achievable and the superior comfort and ergonomics, this type of control system is not widely used because of a glaring flaw: the lack of self-centering and with that, the lack of any type of feedback.

View project log

View all instructions

Enjoy this project?



Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates