Long Range FPV / Surveillance Quadcopter

Overview

This is an ongoing project of mine which started in 2016 in response to a need for a long range, autonomous drone with video capability. After these initial requirements were achieved, I continued to expand the project to incorporate new functionality; an effort which continues to this day.

Changes will be posted here as they are implimented.

Phase V(A) (2023)

As outlined in Phase V, this subphase is dedicated to the 36x optical zoom camera, which was a replacement for the damaged 10x optical zoom camera introduced in Phase IV. The below downlink video from this camera was taken from the same perspective as the previous sample video. It is immediately clear that the 36x camera, gimbal and upgraded control scheme are superior to the 10x camera and servo 'gimbal' from previous phases.


Upgraded 36x camera with gimbal in 'camera man' mode (detailed in Phase V)


Given the 800 TVL resolution of this new sensor (compared to the 600TVL sensor included with the previous 10x unit), the theoretical video resolution limit of conventional analog VTX image transmission has now been reached. In other words, any further upgrades to the FPV camera (sensor) would serve little to no purpose.

Some further improvement is absolutely possible here. Most noticably, the camera roll is slighly off-kilter, which should hopefully be correctable by way of a more detailed gimbal calibration. Another mildly irritating issue is the OSD flickering, which is very likely a symptom of a grounding / bonding issue on the video transmission circuit.

As previously discussed, the analog video transmission resolution limit marks the beginning of the end for the entire analog survaillence suite on this platform. Undoubtably, the next logical upgrade will be a digital transmission system, however more pressing limitations exist on the platform which should be addressed before this switch can occur.

This platform has now reached a very high level of development, and most logical improvements, much like the video survaillence package, will require major overhauls to the platform. In all likelihood, this will come in the form of a completely new UAV platform. Stay tuned!


Is it a miracle? Or a monster?


Phase V (2023)

Per the prediction in the previous phase, gimbal stabilization and an on-screen-display (OSD) were indeed the primary focuses of Phase V. Additionally, several modifications and emergency repairs were made since Phase IV, which ultimately improved system capabilities.

The addition of an OSD was relatively straightforward. A custom microcontroller board was spliced into the video signal line to overlay telemetry collected from the flight controller. This telemetry can then be displayed over top the video feed. Pictured below is the OSD layout containing battery voltage, flight mode, altitude and arrow-indicated direction to point of takeoff.


In-flight OSD output atop video from (now deceased) 10x camera.


I should preface this next section with an announcement of the untimely death of the 10x optical zoom camera featured in Phase IV. Alas, the onboard zoom control board did not take kindly to a reverse polarity 16.8 Volt, 50 Amp load. The death was quick and humane. Despite countless hours of internet sleuthing, I was unable to source a replacment board, and the entire camera was replaced with a upgraded 30x optical zoom unit. While this change was not planned, it did turn out to be something of an upgrade.

Much like finding a replacement zoom board, sourcing a gimbal was also a difficult task, due to the limited number of commercially available models, a global chip shortage, and an overall reduction in the popularity of custom-built platforms in exchange for prebuilt mass-market UAVs.

Ultimately, I opted for an industrial 3-Axis gimbal designed for a FLIR camera. While I would absolutely love to have such a camera (do get in touch if you would like to donate one), the challenge here was to retrofit the gimbal to accomodate the 30x zoom camera. This was accomplished with a re-designed (and sheilded) wiring harness, 3D printed bracket, and some delicate tuning of the gimbal's PID controler to accomodate the larger payload.


Gimbal control in 'camera man' mode (see below)


Most industrial UAV platforms with this capability rely on two operators; a pilot and camera man. Since this was always designed to be a single-man platform, this presented a unique systems engineering challenge, as only one pair of control sticks is present in this system (as opposed to two in the two-operator case). In response to this human input dilemma, I wrote a LUA script for the controller to allow the pitch/roll thumbstick to be switched to steering the gimbal via a two-position switch. This way, a single operator can pilot the UAV and control the gimbal / zoom while at the halt. The result of this 'camera-man' mode is a supremely stable camera with an elegant human interface.

The next phase, V(A), is more of a logical distinction, and will contain some additional details and footage about the upgraded 36x zoom camera, which is the only major difference from this phase.

Phase IV (2023)

Phase IV represents the first upgrade since the resurrection of the project in Phase III.

The preeminent upgrade would be the addition of a 10x optical zoom camera, and pan servo which has vastly upgraded ISR capabilities.

While the new camera functionality is a considerable benefit, the lack of gimbal stabilization or on-screen-display (OSD) will likely become the target of the next phase of integration.


Here is some video of the live downlink. It is very shaky due to the servo-based pan/tilt mount.


Phase III (2022)

Phase III marked the end of the hiatus since Phase II, with my life having sufficiently stablized enough to resume improving upon the UAV.

This phase was also the most comprehensive in terms of overhauling existing design, and included upgrades to nearly every component. Significant repairs were also required as a result of being in and out of storage for several years. The quadcopter was disassembled and rebuilt from the ground up, with the previous Phase II objectives still in mind.

Results were very positive, with all test flights successful. Several long-tem goals were identified, such as another much-desired ISR upgrade, along with a planned upgrade to the now EOL APM 2.6 flight controller.



Phase II (2017)

Phase II introduced a new set of requirements to the platform, including a longer flight time, upgraded camera with tilt functionality, long range command / control link, a more conventional carbon fiber frame, and improved RF compatability / grounding / frequency management.

Unfortunately, my life became extremely busy around the time this phase was being implimented, so development halted for several years, with new objectives being defined upon the project's resurrection in Phase 3. The desired upgrades were identified and purchased, but little integration took place. Some limited test flights were accomplished prior to the hiatus, with strong improvements seen in both flight time and video downlink range.

Per Phase I, a long range UHF command/control link was fitted to replace the generic 2.4GHz radio, as I am a licensed amateur radio operator and can legally utilize the band.

There's a lesson somewhere in here about priorities and life. I'll put my finger on it when I'm done with my drone.



Phase I(A) (2017)

Inspired by the successes of the previous phase, Phase I(A) was a targeted upgrade for the ISR package onboard the drone. The relatively low-resolution camera was swapped out for an improved version, a higher-power transmitter was selected, and a servo added to enable remote camera tilt functionality. The ability to tilt the camera, while simple, greatly improved survaillence, landing, and payload positioning.



Phase I (2016)

Phase I outlined the initial requirements for the UAV, which specified a long range, autonomous platform capable of transporting a GoPro camera, which could fly a survaillence loop around my property. I decided upon an APM 2.6 flight controller atop a custom designed and 3D printed frame. The latter, while a novel idea, wound up being an exercise in frustration.

Between the pitfalls of PLA filament and the length of time necessary to print even a single replacment arm, the benefits of rapid-replacement from 3D printed parts were quickly diminished. Moreover, being the rather careful pilot I am, I simply did not crash enough for it to be worth it.

I later decided to add a 5.8GHz analog video transmitter and statically mounted forward looking camera to enable first-person navigation, which wound up informing a lot of the design decision in the following phases due to the sheer usefulness of the functionality.

Command and control was accomplished with a generic 2.4 GHz radio link, with telemetry being handled by an 800MHz MAVLink radio. These would later be upgraded with higher power transmitters operating on better suited / longer-range frequency bands.


First "high speed" test run- the throttle was software-restricted to 50% output.