November 19, 2012

Remote Exploration of Rocky Planets: from Craters to Smart Dust

Prof John Barker

Report by: Mike Meynell

Prof John Barker

The Flamsteed were treated to a packed evening of entertainment on an otherwise gloomy November evening.

First up, Malcolm Porter gave us a quick overview of objects to view in the night sky this month, focusing on a few objects that are slightly more difficult to view in comparison to those presented at our last lecture.

We were then delighted to welcome Mary Ferguson from the museum, who treated us to a brief update on the Board of Longitude project, whose aim is to digitise the complete archive of the Board of Longitude. The board was formed in 1714 to improve navigation at sea and the archive records one of the earliest examples of state-sponsored science in Britain. Mary would like some help from Flamsteed members, firstly in gaining our views on how people would use the archive and also ascertaining if anyone would like to volunteer to get involved in the project itself.

Mary gave some examples of what is in the archive:

letters from all sectors of society describing instruments and proposals for methods and machines to measure longitude
documentation relating to John Harrison’s encounters with the Board
a detailed logbook from Captain Cook’s second voyage
correspondence with William Bligh regarding the mutiny on the Bounty

A really fascinating topic and one that I’m sure that many Flamsteed members would like to be involved in.

The main event of the night was a lecture on the remote exploration of rocky planets, by Professor John Barker. Professor Barker is a theoretical physicist and Emeritus Professor of Electronics at the University of Glasgow working mainly in the field of nano-electronics. This was a topic that had never been covered before in a Flamsteed lecture as John took us through two main areas:

Looking at impact craters and how they can be identified from images using automatic data processing and minimal sets of data.
How we can explore space in a much cheaper way using ‘smart dust’.

Firstly, John explained how half of NASA’s current budget goes on manned space flight. The Apollo programme cost around $136 billion (in 2007 dollars), whereas any planned manned mission to Mars is predicted to cost around $450 billion. Compare this to the Wilkinson Microwave Anisotropy Probe (WMAP) that enabled the production of a new standard model of cosmology and is best known for mapping the Cosmic Microwave Background radiation, which cost as little as $0.15 billion.

John’s argument is that money spent on the manned spaceflight programme is unsustainable, particularly as unmanned probes and satellites have made most of the major advances in science.

The “Face on Mars”

Firstly, John demonstrated how information could be extracted from seemingly poor quality images to gather information on the topology of a planetary surface. He started with the infamous “Face on Mars” image, which was taken by the Viking 1 spacecraft in 1976. By applying very advanced algorithms to improve the pictures, John and his team have been able to extract information that prove to be a very close match to images which are being produced today by the more advanced probes which have been sent to Mars. Using ‘shape from shading’ algorithms, they have been able to reconstruct 3-dimensional images from the early poor quality images that compare extremely well to images from the latest Mars Express images. So, for the price of a research student, it is possible to create images that compare well with images from a multi-billion dollar satellite! Very impressive stuff.

If we want to automatically explore planets using robotics, we need to find ways of identifying things without human intervention. So, how can surface features such as impact craters be automatically identified using these algorithms? Impact craters are generally circular, even if they hit the planet at an angle, as they are such high-speed impacts. Impact velocities on Earth range from the escape velocity (about 25,000 mph) to over 160,000 mph for objects that are approaching us in the same orbit (effectively, a ‘head-on’ collision).

The first thing we need to search for, therefore, are circles. John highlighted a great website where you can view details on impact craters with coordinates and links to online maps so that you can look at satellite images of them. However, there are lots of circular features on the surface of a planet that are not impact craters, such as natural basins and volcanoes. How can you distinguish between these features and impact craters automatically? It can be done by building a physical model of the crater via mathematical algorithms. John introduced the concept of a Hough Transformation, which is a technique used in image analysis to identify lines or edges in an image. In addition, another technique uses mathematics first identified by Roger Penrose in the 1950s, which takes the equation of the form of structure that is being searched for (for example, a conic section), to give an answer that is the least squares fit to the data. This has the advantage of being a very fast way of analysing data and has proved to be a very accurate method of identifying craters.

So why would we want to do all of this data analysis, when we already have very expensive satellites capable of producing high quality images? John asked what would happen if funding for space exploration was radically cut. In our current ‘age of austerity’, all public spending is under pressure. Are there ways of doing space exploration ‘on the cheap’, with poor quality data but using new technologies and algorithms to analyse the data? This brought John to describe the concept of ‘smart dust’.

Smart dust are swarms of very small computer chips, around 1-2mm across, containing billions of components. The smart dust could be dropped on a planet by a ‘mother-ship’. If they were equipped with small CCD sensors, even with high noise and poor dynamic range, they would give equivalent ‘poor quality’ images like the Viking data, which John had already shown could be processed to give a very accurate analysis of surface features.

The major problem with smart dust is how do you move it across the surface of a planet? John realised that this could be done by adapting the shape of the smart dust by encapsulating them in a shape changing electro-active polymer, which can distort as tiny voltages are applied to it. As the smart dust particles are approximately the size of sand, they will move in the same way with the wind. The key process in the movement of sand across a surface is known as ‘saltation’, where particles ‘hop’ across the surface.

Mars, of course, is covered in sand, so these smart dust particles could be used to explore the surface. With wind speeds on Mars much higher than the Earth and gravity much lower, it should be possible to explore large areas of the surface of Mars using the wind to transport smart dust from one place to another.

However, you would want to keep the dust together so that the dust particles can communicate with each other. This is where the shape changing comes in. A particle can change shape to increase its drag, thereby travelling further in the wind. If the particle is going in the right direction, it stays the same shape. If going in the wrong direction, they make themselves smoother, thereby dropping to the ground where they wait until the wind picks up again and checking again if they are going in the right direction. In this way, the particles can be kept in roughly the same formation.

Getting the data back from the smart dust can be done by adjusting the phase of the electromagnetic signal, sending a powerful signal back to the mother-ship in orbit around the planet.

In conclusion, John stated that it was possible to have low cost techniques for gathering data in space exploration, implementing all of this with ‘swarm robotics’, such as smart dust.

An absolutely fascinating lecture that kept the Flamsteed audience enthralled throughout. Our sincere thanks to Professor John Barker.