Unmanned System Data Protocol and Format


This week’s blog will focus on DJI’s Inspire 2, a professional quality quadrotor Unmanned Aerial System (UAS) designed for filmmakers.  The UAS is capable of speeds of up to 58 mph and has an endurance of 27 minutes through the use of dual, redundant, battery packs (Inspire 2, 2017).  In addition to an integrated first-person view (FPV), 2 axis, 4K camera built into the nose, the aircraft can support a variety of 360 degree rotatable gimbled film cameras, including DJI’s top of the line Zenmuse X7, capable of recording in 6K CinemaDNG at 30 frames per second (fps) or 5.2K Apple ProRes video formats at 30 fps; both formats are capable of 60 fps rates at reduced quality (Inspire 2, 2017; Zenmuse X7, 2017).  The Zenmuse X7 camera is also capable of capturing DNG RAW 20 fps (burst mode) still photos at 24 megapixels (MP), as well as the commonly used JEPG format (Zenmuse X7, 2017). 
 DJI. (2017). Inspire 2 [digital image]. Retrieved from https://www.dronethusiast.com/wp-content/uploads/2017/01/dji-inspire-2-pic.png

The Inspire 2 is capable of recording and simultaneously transmitting imagery captured by the gimbled film camera.  Seamless image processing and storage of recorded images is accomplished through the use of on onboard processor called CineCore 2.0 and an integrated solid state drive, called CINESSD (Inspire 2, 2017).  The use of an onboard processor is critical as the high quality CinemaDNG format has bit rates of 4.44 Gbps; the burst rate for DNG RAW still images is higher at 4.70 Gbps (Zenmuse X7, 2017).  Recorded video footage is processed by CineCore and then routed to the CINESSD (for CinemaDNG or AppleProRes formats), or to a Micro SD card for lower quality 4K H.264 and H.265 video formats (MOV, MP4) (Inspire 2, 2017).  The CINESSD and Micro SD storage devices both utilize the FAT32/exFAT file format system, as this commonly utilized formatting method eliminates the need for additional software to transfer video and images files from the storage devices upon landing.  Maximum onboard storage capacity is 480 GB for the CINESSD and 128 GB for the Micro SD card (Inspire 2 User Manual, 2017). 

 
DJI. (2017). Workflow [digital image]. Retrieved from https://www.dji.com/inspire-2?site=brandsite&from=nav
Footage captured by the onboard FPV and film cameras is also transmitted to the operator on the ground.  While it is possible for a single operator to record footage and operate the UAS, the UAS is ideally operated by a dedicated film camera operator and dedicated pilot.  This method of operation enables the camera operator to focus on the quality of shot while the pilot maneuvers to avoid obstacles and obtain the desired perspective.  Seamless video footage is broadcast to the ground operator/s at 1080i, at up to 4.3 miles over 2.4 GHz or 5.8 GHz frequencies utilizing DJI’s Lightbridge video transmission technology (Inspire 2, 2017).  This live imagery can also be directly broadcast to media outlets by connecting the UAS remote controller to a video input on a news station’s satellite truck (Inspire 2, 2017). 

The Inspire 2 utilizes a host of onboard sensors to maintain attitude and altitude in flight and detect and avoid obstacles.  Aircraft positioning and attitude are maintained through the use of GPS and redundant internal measurement units (IMU) and barometers (Inspire 2, 2017).  Obstacles avoidance sensors include forward and downward facing pairs of stereo vision sensors which enable the UAS to detect obstacles approximately 100 feet away.  This enables the UAS to return to home autonomously at speeds up to 34 mph without risk of obstacle collision (Inspire, 2017).  These sensors are also utilized for hovering and position keeping when a GPS signal is not available.  An upward facing infrared sensor is utilized to detect and avoid obstacles up to 16 feet above the UAS (Inspire 2, 2017).  Inputs from the onboard sensors are fed into the intelligent flight controller for navigation and avoidance of detected obstacles. 


DJI. (2017). Sense and Avoid [digital image]. Retrieved from https://www.dji.com/inspire-2?site=brandsite&from=nav

Power for the onboard sensors, cameras, and rotors is provided through the use of dual redundant batteries.  The use of dual batteries enables the UAS to execute a safe landing utilizing one battery in the event a problem occurs with the other battery (Inspire 2, 2017).  Each lithium polymer (LiPo) battery has a capacity of 4280 mAh, a rating of 98 Wh, and an output voltage of 22.8 V for onboard motors, sensors, and cameras (Inspire 2 Specs, 2017).  Maximum battery life of 27 minutes is obtained utilizing the optional X4S film camera; use of the more powerful X5S camera decreases endurance slightly to 25 minutes (Inspire 2, 2017).  Battery endurance utilizing the top of the line X7 camera is not available. 

One alternative data treatment strategy to improve the Inspire 2 would be the use of a lossless video compression algorithm to increase the current 1080i quality of live video streams for use by media outlets or filmmakers.  The ability to downlink higher quality imagery would reduce the need to store recorded imagery on the internal CINESSD, potentially eliminating instances where the available space remaining on the hard drive, rather than battery endurance, was the limiting factor to flight time.  For example, the currently utilized 480 GB CINESSD drive can only record 108 seconds utilizing the highest quality CinemaDNG recording setting which has a bit rate of 4.44 Gbps.  The ability to downlink high quality imagery could also reduce the size of the storage drive required, potentially allowing for larger batteries which would increase flight endurance. 

References: 
Inspire 2. (2017). Retrieved from https://www.dji.com/inspire-2?site=brandsite&from=nav.
Inspire 2 Specs. (2017). Retrieved from https://www.dji.com/inspire-2/info#specs.
Inspire 2 User Manual. (2017). Retrieved from https://dl.djicdn.com/downloads/inspire_2/ 20171013/ INSPIRE_2_User_Manual_EN.pdf.
Zenmuse X7. (2017). Retrieved from https://www.dji.com/zenmuse-x7?site=brandsite&from=nav.

Comments

  1. T.J. AdamsDecember 13, 2017 at 4:40 PM

    Good overview of the Inspire 2. I was able to fly an Inspire 1 for the first time today. That combined with your detailed post were beneficial in understanding the differences between the Inspire 1 and 2. The extra storage capacity with the CINESSD in addition to the MicroSD card is a nice upgrade from earlier systems. I definitely agree that a separate pilot and camera operator are needed for better footage. Hard to do both effectively!

    ReplyDelete
  2. UnknownDecember 13, 2017 at 8:29 PM

    Great write up,
    Since i purchased my first quadcopter a couple months ago, im very intrigued with fast flying ones. I am looking to purchase the BeBOP 2 power by Parrot. When things are moving that fast and at low level flight it is very hard an you stated to remain consistent. Its not like a UAV flying by at over 10,000 feet. Having a separate sensor operator and pilot would be ideal. That is how the professionals do it, and it works wonders.

    ReplyDelete

Post a Comment

Popular posts from this blog

Unmanned Systems Maritime Search and Rescue: IVER3 AUV

Image
OceanServer. (n.d.). IVER3 [digital image]. Retrieved from www.iver-auv.com/AP_003_Iver3.pdf The US Navy deployed three IVER3-580 Autonomous Underwater Vehicles (AUV) to Argentina during the recent search for the Argentine submarine A.R.A. San Juan, missing in the South Atlantic Ocean since November 15th (US Navy, 2017).   The IVER3 AUV can operate at depths of up to 100 meters at speeds of up to 4 knots for up to 14 hours (Overview, 2017).   Active search sensors include a side scan sonar, which is utilized to obtain an image of the ocean floor to detect objects, and a swath bathymetry sonar to map the depth of the ocean floor below the vehicle (IVER3, n.d.).   Navigational sensors include GPS for fixing the vehicle’s location on the surface, a doppler velocity log (DVL) and inertial navigation system (INS) for determining vehicle location under the surface, and a forward-looking echo sounder in the bow for collision avoidance (IVER3, n.d.).   The vehicle can also be e
Read more

Combining exteroceptive sensors aboard a UGV

Image
This initial blog post will discuss the use of exteroceptive sensor systems aboard the Daredevil PackBot unmanned ground vehicle (UGV) to navigate in all weather conditions, as outlined in Yamauchi’s paper titled “Fusing Ultra-Wideband Radar (UWB) and LIDAR for Small UGV Navigation in All-Weather Conditions.” Yamauchi’s research into the use of UWB radar and LIDAR builds upon previous research that showed UWB radar was capable of detecting objects “in a snowstorm, through dense fog, and through sparse foliage” (2010).   However, the use of radar is not as accurate as LIDAR or visual navigation techniques, reducing the speed at which the UGV can operate.   The use of a calibrated max filter algorithm addressed issues experience differentiating between ground clutter returns and obstacles, but remained unable to match the accuracy of LIDAR or stereo vision sensors.   Yamauchi, Brian. (2010). Daredevil Packbot [digital image]. doi: 10.1117/12.850386 Advantages to the incorporati
Read more

Sense and Avoid for Small Unmanned Aerial Systems

Image
Aerotenna’s uSharp Sense-and-Avoid Radar is a commercially available sense and avoid solution for small Unmanned Aerial Systems (sUAS).   The uSharp radar utilizes microwave radar in the K-band (24 GHz) frequency range to detect objects in the horizontal plane within a full 360 degrees (Aerotenna Technology, n.d.).   The use of microwave radar as opposed to commonly utilized sensors such as electrooptical (EO) and infrared (IR) cameras or Light Detection and Ranging (LiDAR), enables a sUAS equipped with uSharp to sense and avoid other objects in all weather (heavy rain, fog, snow, clouds) and visibility conditions (night or low light levels) (Aerotenna Technology, n.d.).   The radar consists of eight antennas arranged in an octagonal formation and has a highly accurate range resolution of five centimeters and a response rate of 90 Hertz (Aerotenna, 2016).   It also has a range of 0.5 to 120 meters and can detect both moving and stationary objects (uSharp, 2017).   The ability to de
Read more