Over the last few weeks drones have been swarming our headlines. From their role in opening and closing the Winter Olympics to strutting their stuff above the catwalk at the Milan Fashion Week, it is clear that the drone industry is taking off in new and unexpected ways.
Research into drones has previously been predominantly driven by the military. However, due to the reduction in size of the processors needed to control the drones (something in part driven by mobile phone technology), drones can now be made much smaller and lighter, which facilitates a greater range of applications. It also helps to reduce the costs associated with research. As a result, the range of fields in which drones (most commonly unmanned aerial vehicles (UAVs)) are used is growing at an incredible rate and technology needed to facilitate these new applications is keeping pace. Where previously they were predominantly deployed by military forces, they are now even being developed to plant trees, prevent poaching and deliver blood.
Drones come in an extraordinary variety of shapes and sizes and use a wide range of technologies. Drone development therefore involves many different disciplines ranging from aeronautical and mechanical engineering, to electronics, software, materials, communications and power systems. An attempt to describe all the developments in these fields would be futile, so we’ll focus on two notable companies working in this sector.
A tale of two companies
Let’s take a look at Intel and DJI and explore their recent IPdevelopments. The former’s record-breaking light show at the opening ceremony of the Winter Olympics involved a swarm of 1,218 drones flying together to create a breath-taking visual display (also known as a digital firework display and definitely worth a watch). The latter have a global market share of the civilian drone market of approximately 70 per cent [1]and are widely regarded as the world’s leading drone company.
In an attempt to understand Intel’s involvement in the drone sector, a search[2] was performed on Espacenet for patent applications and granted patents (IP) relating to drones with Intel as the applicant. The results were then categorised. The same search was then performed but with DJI as the applicant and the results compared. Whilst the data set is not vast, the results show some differences, as would be expected, between the two companies:
Intel intelligent
Whilst Intel have indeed just set a world record, the drones behind the display are in fact remarkably simple pieces of kit. Where Intel have excelled is in their communication systems and data processing as well as the applications which stem from these. A recently published patent application describes a universal controller for robust trajectory tracking in multirotor unmanned aerial vehicles, outlining a control system that may well have been used to help them to achieve this record. This includes a sensor system to give position and orientation data and a flight control system configured to compute error values between desired trajectories and actual trajectories and generate control signals to counteract the computed error values in each dimension of 3D space [3]. Sound complicated? It is. What this robust control facilitates is the ability to change payloads, conditions or even vehicles without any requirement for reconfiguration of the controller for autonomous operation. This sort of dynamic controllability will be extremely valuable in the future as hardware is improved and versatility of systems to the range of applications and uses for drones increases.
The range of potential uses for drones is indeed enormous; it is not simply for digital firework displays that Intel is developing this technology. A patent application published earlier this month demonstrates Intel’s ambition to use drones to assist in emergency responses by navigating in advance of an emergency vehicle[4], an application for which their object detection and recognition module[5] may come in handy. What can be seen from Intel’s recently published applications is that they are sticking to what they know, with a focus on communication systems[6], control systems[7] and data processing and collection[8], [9], [10]. Intel even envisage using drones to generate 3D maps of data centres[11]: data about data.
DJI dominates
What has underpinned DJI’s astonishing growth and success is the booming civilian “drone enthusiast” market. However, whilst this does crucially involve data collection (in the sense of taking photos and videos) the company has first and foremost been a hardware company. Indeed, their first “drone” patent application filed in 2011 actually relates to a camera gimbal system to be mounted on a drone to improve the quality of the pictures and videos (data). DJI are widely renowned for producing low-cost, high-quality and easy-to-use drones. Their users want drones that fly longer, faster and further, are more stable in the air and have a ton of cool, easy to use features such as object tracking and waypoint planning for easy use. This is reflected in their IP which, as can be seen above, shows a focus on airframes and other hardware, propulsion and power systems and control systems.
Applications
It naturally follows that DJI have a wealth of IP aimed at enhancing user experience to help ensure they maintain their edge in the consumer drone market. Recently published patent applications describe methods for: safely hand landing drones[12]; flight planning[13]; managing flight-restriction regions[14]; geofencing[15]; and interactive UAV video broadcasting[16]. However, more recently DJI have been targeting the ever-growing industrial sector with new hardware such as the Matrice 600and the Agras MG-1S, both of which are used with their new Agriculture Management Platform (underpinned by their IP[17]), as well as new sensors and systems for multispectral analysis and thermal imaging. The range of potential applications is indeed vast, with DJI filing patent applications for drones that clean walls[18] and even drones for walking and training dogs through the administering of positive and negative stimuli[19] (see figure below).
Hardware
A patent application filed as far back as 2012[20] depicts the first of the series of “phantom” drones that defined DJI in their early stages. This application describes separating the onboard sensors from other high voltage electronic components in order to reduce interference to sensor readings (thereby improving reliability of operation). The phantom drone is still what the majority of the world think of when they think of drones and is the drone of choice to feature on any BBC article discussing airports! However, since 2012, DJI have stepped things up a notch. Their more recent patent applications include UAVs configured for multi-orientation flight[21]; UAVs with tilting rotor mechanisms[22]; UAVs with legs for walking when on the ground[23]; foldable UAVs[24], including the IPbehind their hugely successful Mavic drone[25]; and automatic docking, recharging[26] and battery swapping[27] stations.
It is clear that there are a number of these systems that can be put to good use in terms of improving the range of products available to drone users: folding drones (such as the Mavic) increase portability; tilting rotors may improve camera stability; and multi-orientation flight and hybrid vehicles with legs are fun “new” features that DJI’s loyal fan base will be only too happy to pay for. However, systems such as the automatic recharging devices are more likely to be of benefit in the industrial sector where automation is very important in improving the efficiency of new systems. When (if) the drone enthusiasts have had their fill, it will be industry that continues to drive DJI and the drone market forward.
Control systems
Whether it is to optimise for industrial application or to make next year’s Christmas present extra special, what all of these drones and applications do require are reliable, versatile and well-tuned control systems. DJI’s extensive IP in the field of control systems reflects this need and includes: methods for obstacle avoidance that utilise circular LiDAR like systems[28]; image-based depth perception systems[29]; automatic target tracking systems[30], [31]; image-based hovering control methods[32]; autonomous flight planning[33]; return to home function[34]; and target tracking control[35].
Power and propulsion
One of the key problems that still faces drone users is poor endurance. Multi-rotors in particular are power-hungry and the short flight times resulting from this are one of the most limiting factors of operation for many applications. This is just one of the reasons for DJI’s battery swapping and automatic recharging stations as well as for their additional IP in the fields of propulsion and power systems. For example, DJI aim to make more efficient systems (that fly longer) with: systems designed to recover wasted heat[36]; systems for recovering and recycling motor power from decelerating propellers[37]; and propellers optimised for reduced resistance[38].
Back to the future
If there is a trend to be spotted, then it is a trend towards automation and autonomy of operation. Whatever speciality one particular company may have though, for any of these applications to be realised, solutions from all of these fields are required. By sticking to what they know, DJI and Intel have achieved some remarkable results both technologically and through resounding business success. All it now takes is the right people to put these systems together and the sky really is the limit.
For further information, please contact Theo Carter or your usual Kilburn & Strode advisor.
[1] https://www.ft.com/content/0d87a148-7d7f-11e7-ab01-a13271d1ee9c
[2] Search on Espacenet performed 22 February 2018 for “fly or flying or drone or uav or swarm in the title or abstract AND intel or DJI as the applicant”
[3] WO2017/111971
[4] WO 2018/022021
[5] US 2017/0102699
[6] US 2016/0378109
[7] US 2017/0285633
[8] US 2017/0351933
[9] US 2017/0330040
[10] US 2017/0090483
[11] US 2017/0247108
[12] US 2017/0315562
[13] WO 2018/006216
[14] WO 2017/084031
[15] WO 2016/154936
[16] WO 2018/014338
[17] WO 2018/000399
[18] US 2017/0158329
[19] US 2017/0215381
[20] WO 2014/075609
[21] WO 2018/010097
[22] WO 2017/219255
[23] WO 2017/219295
[24] WO 2017/205997
[25] WO 2017/143645
[26] US 2017/0253349
[27] US 2017/0129464
[28] WO 2018/014252
[29] WO 2018/006296
[30] US 2017/0322551
[31] WO 2017/147789
[32] WO 2017/166002
[33] WO 2017/143588
[34] WO 2017/096548
[35] WO 2017/096547
[36] WO 2018/014336
[37] US 2017/0334570
[38] WO 2017/128742