Matt Hubbard, of Anthony Best Dynamics, returned to give another talk about driverless car testing. He briefly summarised the robotic steering and pedal controls that were the basis of the 2011 talk and how they had been developed to maintain a vehicle within 2 cm lateral and 10 cm longitudinal of its target position.
However the bulk of his talk related to newer products, the Guided Soft Targets. The original concept was to produce a 'soft' guided vehicle which could survive 50g accelerations yet not damage any conventional vehicle that it might drive into.
These targets were needed to trigger collision avoidance systems on test cars, so were made to look like 'real' cars visually and as radar targets.
The latest development has been to produce the Guided Soft Target, a vehicle that is designed to be collided with yet to do no significant damage to the test vehicle.
Matt described the concept, originating from California, of a very low-profile aluminium frame carrying batteries, control electronics, drive motors and hydraulic systems. On top of this frame foam panels are slotted together to replicate the appearance of a standard saloon car. A real car hitting this target dislodges the foam panels and drives over the GST's chassis. The GST can survive 50-100 of such collisions.
Matt described the mechanical arrangements of the vehicle, two steering wheels at the front and one powered wheel at the rear. Two electric motors are needed to drive the GST because of the restricted height and hence motor diameter. Bicycle disc brakes are fitted to each front wheel and drive motor, though with thicker discs to add thermal mass.The GST has moved from using sealed lead acid batteries to Lithium Iron Phosphate cells. These later cells give improved energy density but require more sophisticated charging arrangements compared to Lead Acid. Although safer than Lithium Ion batteries, it is proving difficult to ensure that airlines will accept them. The GST is controlled via a Wi-Fi link. Two on-board controllers 'watchdog' each other. One controller essentially executes the test task while the other brings the vehicle to a halt in the event of a control failure. This second controller manages a secondary braking system.
Matt closed his talk with a brief description of the company, growing from 18 employees when he started to 75 now. The company has floated on the stock exchange, providing new development capital and allowing the original owners to liquidate some of their equity. He expressed his disappointment with the quality of some of their recent engineering applicants who were lacking in basic knowledge.
On the rare occasions when the centre repeats a subject there must always be a concern that the content too will be repeated. Fortunately engineering always moves on and so it proved in this case. Whilst the first talk had been about ensuring repeatability when driving 'real' test cars this talk was about providing a 'target' environment for them. The low-profile GST is an interesting concept; how does the matador get out of the way of the bull? Let the bull charge over him! Matt said that their best employees had a practical side to their background and I'm sure that struck a chord with most of the audience. My own childhood was influenced by engineers of one of Britain's most inventive marques and they were very-much hands-on, literally putting their lives on the line when it came to testing their own designs. They would have been fascinated how new tools, such as the subject of to-night's talk,allow their successors to develop ever safer vehicles without the risks of the past.
