Image found at and rights belong to http://www.intropsych.com/
This may come as a surprise (or not) that I grew up in a household with Psychiatrists; my parents. Before you jump to the ‘did your parents analyze you’ and all the media stereotypes, my parents are and were parents and people first and foremost and didn’t try any of that rubbish on me growing up. However, what did transpire in our household was lively discussions and debates, lessons and exposure and stimulation both creatively and technically. I can honestly say my parents kept a wonderful balance of never closing any door to me while always advising and trusting me through years. What I value the most, however, has always been the discussions. In the last few years I can remember two that stood out; one on a bench upon the Eiffel tower (with Mom & Dad) and the other in a coffee shop with my Dad. More than the subject matter, I can always remember the sense of connection and happiness knowing that I can always speak to my parents openly and still learn from them, no matter how far I think I’ve come.
Having Psychiatrist parents, the subject of the brain has always been in rotation, albeit in different forms. We’ve discussed behavior, structure, chemistry, science and on and on. What is fascinating, to me, is that despite the infinite variation of expression, the core structure is ultimately the same. And with it’s complexity, simplicity and ability, even with our greatest technology are still not capable of matching it. The human brain weighs in at about 3lbs, has near 1 quadrillion connections, never shuts down, is powered by energy derived from naturally occurring food sources, and while it is said it consumes 20% of our total energy usage it regulates an incredibly sophisticated machine.
Thusly, over time, I have learned more about the structure of the human brain, and in it’s most basic it has several core ‘lobes.’ The Frontal, Parietal Lobe, Occipital and Temporal lobe, each with their own specific (and some overlapping) purposes. The Frontal lobe handles reasoning, planning & problem solving. The Parietal, movement, orientation, recognition and stimuli. The Occipital lobe is nearly exclusively for dealing with visual processing, while the Temporal lobe handles auditory, memory and speech matters.
From here, I realized that where I wished the future of drones to head correlated with the logical structure of the brain. I began to see the brain as a set of processors handling different core functions and then correlating them to one another. In fact, this exists today, look under the hood of your average computer or smart device and you’ll see many of the same features. Core processors handle the reasoning & problem solving, apps the planning. Gyros & accelerometers handle movement and orientation, your screen and graphics chip the Occipital while RAM and DSP chips taking the role of the Temporal lobe.
Extracting this in to drones, however, requires some different parts. If we take a deeper look in to the brain, there are areas with more specific functions. The Cerebellum coordinates fine motor balance & muscle movement, the spinal cord is a singular core data relay, the Thalamus acts as the call center and the Hypothalamus regulates hunger, thirst & temperature control.
When I put it all together I see drone system structured as such:
Cortex made of the Frontal & Parietal lobes: This will the a new processor & computer – capable of complex simultaneous processing to solve and deal with multiple inputs at any given time. It must be low power and part of a computer system able to run apps allowing users to complex actions simply. Furthermore a set of gyroscopic and acceleration sensors should provide further depth to information. I’m thinking about Parallel processing for this purpose.
Cerebellum: Nearly if not equally important as the Cortex in this application is the cerebellum. Controlling fine motor control is up to the Autopilot or Flight controller in most drones. Empowered by the decreasing cost of microchips, they have become more sophisticated and faster than before. Today, they are powered by either 16bit or more recently 32bit processors. Despite this, they have never combined with a more sophisticated system, say like Android. The reason for this, as I understand it, is they require a Real-Time Operating System or RTOS; it operates quickly and predictably on a narrow set of parameters. That said, I don’t think the Autopilot should be folded in to, but remain a whole separate component to the computing system with it’s own set of sensors, however it should be able to stream status information to the ‘Cortex’ and receive commands as well.
Occipital Lobe: This will be the visual center – utilizing algorithms and graphics chips to do realtime object recognition at high framerates as well as act as the video card for the computer system. The video stream, in this conception, would hit two sets of video processors first the object recognition, then the video graphics, then pass through to the computer. The video graphics in this application could provide a more sophisticated OSD (or On Screen Display) than current FPV users are used to.
Thalamus: The thalamus represents the call center and handles the incoming and outgoing signals of the cortex. As I realized that I want to incorporate several subsystems in to this new platform, the need for a means of unified information handling and sharing arose. I’m still investigating methods used in other applications, but what has come to mind as of late is the idea of a computer system with it’s various PCI expansion ports, or PCMCIA in the old laptop days. A modular system which can accept new upgrades to the system and shares a communication pathway for the Autopilot and the Visual Processing system to the Cortex.
Spinal Cord: The Spinal Cord is the last piece in the action & motive side of the equation for me. It symbolizes the relay of information from the brain to the body. In the last year or so, there has been continuing development of a new speed controller. Speed controllers, for the uninitiated, are the link between power and the motor & control and the motor. The speed controller outputs power to the motor as dictated by the flight control system. This is typically done with a PWM signal (or Pulse Width Modulation), which has limitations in resolution and speed when compared to other systems. And these new speed controllers utilize a system called I2C which allows for both a simplified set of connections between the controllers and the flight system, and offer higher resolution at faster speeds, well out of the capabilities of the current motors. I see this as a future-proofing step which may offer the opportunity of simplifying and reducing the size of the speed controller circuitry.
Temporal Lobe: The memory and audio portion of the brain still eludes me a bit. Memory is obvious, serving as both RAM & possible HDD for video capture etc; however the purpose of audio on a drone is still up in the air. I can see a speaker for auditory status alerts ala Aibo (but much less cute), but is hearing useful? I tend to use echo-location a bit more than the average person; when dropping a single screw I can generally track its whereabouts even if it goes out of visual range. So, I thought this insight may suggest that if the drone could ‘hear’ it may be able to estimate it’s proximity to objects as a redundant system, but I’m still investigating this.
Hypothalamus: And finally, I imagine that with additional computing power and more silicon, heat maybe an issue, as well as battery life. I do think that battery life should have it’s own circuit which can monitor it and share it’s status with the rest of the system. 3DRobotics created such an accessory which was quite useful (http://store.3drobotics.com/products/apm-power-module-with-xt60-connectors). Their power module acts as a regulator and an information relay about battery stats to the rest of the system and I expect others to follow their lead, as should I. But, cooling is another issue. I think drones of the future shouldn’t be limited in their use by environmental conditions, so I don’t think big gaping apertures are the answer. I still need to do some thinking on this matter, but there are active & passive cooling systems, air and liquid. Liquid is quite interesting as it most represents biology. However, given that liquid cooling in most computer applications, that I know of, is both stationary and heavy, I have to question its usefulness in this application.
And while all the pieces aren’t there yet, I see the potential for the development of something entirely new, more sophisticated and capable than ever before. I know this line of thought isn’t new, scientists have been chasing this dream for a long time, however I think my approach maybe more practical and based on available parts. I’ve begun to slowly put together this system and while its in its infancy, the prospects are exciting. You see, I never forgot the conversations I had with my parents. I know that some part of them was crestfallen when I didn’t follow their career pathway, despite my love of psychology, but I see the world in a slightly different fashion and I still want to take the knowledge and experience they have shared with me and use it in my own way, wish me luck.
Horus Drones 