Where should you put that sensor?

Sensors don't always do what they should

Almost every modern car, truck, tractor, harvester, and self-propelled sprayer has inbuilt processors and sensors that monitor and adjust machine performance. Sensors can help reduce vehicle emissions, improve fuel consumption, improve traction, and enhance machine efficiency. Sensors have been added to combine harvesters that will allow them to sense the crop in front of them, and adjust machine settings before it reaches the crop. Crop sprayers, with section and nozzle control systems, can now make use of cameras to detect the presence of a weed and activate the nozzle. The next generation of seeding equipment could have sensors mounted on every sowing tyne to adjust sowing depth via pressure exerted by the press wheel.

The level of sensing promised to agriculture is stunning. The number of curse words uttered by a farmer, when these sensors disable a machine due to a blown fuse, faulty wire, or build-up of dust in a hard-to-reach place is equally stunning. Farmer’s ability to quickly repair sensors, and enable a machine to continue to function, when the job must get done has diminished. Downtime becomes costly, and frustrating, as the farmer must wait for a technician to arrive and service the machine. The question therefore needs to be asked: Should we simplify the machines, and sense the crop with other systems?

For example, a high-performance UAV equipped with a high-end sensor, that is superior to that mounted on any spray unit, could remove the need for complex wiring and reduce the level of on-board compute from the spray unit. Not only would the spray unit be cheaper to purchase, the sensing system wouldn’t disable the sprayer entirely. It could still function without the UAV. Furthermore, with the UAV scan, you may decide not to worry about activating selective spraying.

This concept could be developed further, with advanced UAV mounted sensors. Soil moisture status may assist modify seeding regimes, and even help assist understand where crop emergence problems may occur. Advances in machine learning and image processing are required to realise this potential.

Like all technology though, viable, multi-tasked UAV platforms for agriculture do not yet exist. Processing technology, and compute could quickly become problematic, but, it is possible to challenge the notion that every sensor and activator must reside in the parent machine. Simpler, cheaper machines, that can be augmented with off-machine sensing systems could help growers break what might appear to be built-in obsolescence that challenges some machines, where the sensors, rather than the gearbox and engine have rendered the machine obsolete.