Manifesto for a Universal Wireless Sensor Network

Wireless sensor networks will not talk to each other

If we are to have a vast network of wireless sensors, we need to convince sensor manufacturers to create sensor modules. Manufacturers are encouraged to copy and add enhancements to this platform, and still produce networks using their own, proprietary, protocols.

What we would like to do in this project, is to formulate a protocol and open source hardware for a ubiquitous wireless sensor network. We are proposing a protocol which may be employed by LoRaWan, LoRa, WiFi and any other network. Gateways will enable communication between networks, but the sensor data protocol will be universal.

Measurement Units

One of the main issues standing in the way of a universal sensor network, is the non-universality of measurement units. This has impacted even projects such as the Hubble Telescope. Unit conversion is sometimes fraught with mistakes.

We propose to use the Unified Code for Units of Measure, found here http://www.unitsofmeasure.org, as a starting (and possible, an end point, as it is very comprehensive). You are encouraged to read this specification, as it gives a wonderful perspective on measurement, starting from earliest history. This specification also allows for new measurement units to be created using the BNF specification in the diagram below:

Common sensor measurement units

For the sake of brevity, we list only the units we believe to be the most commonly used. Readers are encouraged to read the addenda to the specification, particularly the section C.3 ALPHABETIC INDEX BY KIND OF QUANTITY

Temperature, Pressure, Humidity

name	kind of quantity	c/s	c/i	definition
kelvin	temperature	K	K
Celsius	temperature	Cel	CEL
Fahrenheit	temperature	[degF]	[DEGF]
pascal	pressure	Pa	PAL
relative humidity	humidity	%RH	%RH
absolute humidity	humidity	AHD	AHD	g/m3
ampere	electric current	A	A	C/s
volt	electric potential	V	V
millivolt	electric potential	mV	mV
microvolt	electric potential	uV	uV
ohm	electric resistance	Ohm	OHM	V/A
weber	magnetic flux	Wb	WB	V.s
siemens	electric conductance	S	SIE	Ohm-1
tesla	magnetic flux density	T	T	Wb/m2
henry	inductance	H	H	Wb/A
lumen	luminous flux	lm	LM	cd.sr
lux	illuminance	lx	LX	lm/m2
second	time	s	S
minute	time	min	MIN	60s
hour	time	h	HR	60min
day	time	d	DAY	24h
where c/s = case-sensitive, c/i = case-insensitive

Usage

The usage section will be expanded and changed as more collaborators join the project. If you have input, suggestions, or just want to join for updates, please contact us at join us.

Protocols

1. MQTT (The Things Network https://www.thethingsnetwork.org/)
3. Us....

1. MQTT

2. ......

3. Proposed Sensor Data Packaging

To conserve as much battery power as possible, data packages have to be as small as reasonably possible. In order to make data packages ubiquitous and human-readable, we propose an ASCII data format as well as a binary format. We respectfully request community input on this.

ASCII

We propose a data package consisting of:

datetime	number of bytes in read	reading	units
ISO 8601:2000 YYMMDDTHHMMSSZZZZZ	bytes in data	ASCII	as above
181231T235959-0200

It is entirely reasonable to present time as a binary number, and/or an offset, but in the interest of readability we propose using YYMMDDTHHMMSSZZZZZ. We propose to use UTC for all sensor packages, and we request community input regarding this.

A sample sensor reading looks like this after formatting:

datetime	number of data bytes	reading	units
180312T230530-0600	5	-12.0	CEL
alternatively:
180312230530 (all readings in UTC)

Multiple readings in a data packet

In the case of modules containing multiple sensors, and where the datetime information is shared, we propose the following:

datetime	number of bytes in read	number of reading packages to follow	unit	termination
180312T230530-0600	0	3	PKG	linefeed (\n)
	5	-10.0	CEL	\n
	4	22.5	%RH	\n
	6	-12.46	V	\n

Binary

To follow.

References:

https://www.eclipse.org/uomo/ http://www.unitsofmeasure.org/trac https://jcp.org/en/jsr/detail?id=256

Interesting Reading:

Toward a Robust Sparse Data Representation for Wireless Sensor Networks