The Ambient REP publishes the RAI interface for a number of resources (or sensor nodes) in the Ambient Legacy WSN. It also provides access to the WP4 innovations that are integrated in the Ambient Legacy WSN. For this purpose the Ambient REP communicates with the Ambient Middleware API for sending requests, for getting responses and for obtaining subscribed sensor updates. The Ambient REP will convert request from SENSEI applications into API requests on the middleware and will convert API responses and updates for registered resources to RAI responses.

Table of Contents

  1. WP4 innovations available through the AmbientREP
  2. Architecture Overview
  3. Current features
  4. Open issues
  5. Planned features
  6. Requirements
  7. How to build AmbientREP
  8. Obtaining other nodes and software
  9. How to test it

WP4 innovations available through the AmbientREP

  • Self calibrated localization: Real-time localization of mobile nodes in a dynamic environment with self calibration. Benefits are:
    • adapting to the environment and changing conditions on-the-fly without taking extra measurements
    • automatically calibrate nuisance parameters (path-loss exponent and reference RSS) from localization measurements, instead of a pre-calibration phase which is required in most other RSS-based methods
  • Activity monitoring (as part of activity recognition and monitoring): The activity level (or IMA value) is calculated on inertial sensor nodes worn by persons, to indicate the amount of movement of that person. Doing this calculation on the nodes saves bandwidth since the raw inertial sensor data would consume much more bandwidth. Additionally, applications do not have to deal with raw sensor data which is often sensor-specific. Other benefits of activity recognition and monitoring are:
    • scales well with number of sensors/actuators
    • high accuracy and low false alarm
    • distributed and real-time
    • low memory and processing power are required
  • Outlier detection (as part of outlier and event detection): identifying incorrect measurements within the sensor network and correcting them either by removing them or use them for further analysis helps in saving energy and ensuring data quality. Benefits of outlier and event detection are:
    • robustness against various failures, i.e., sensor, node, or communication link failure
    • adaptive and able to learn
    • low memory and processing power are required
    • high accuracy and low false alarm
    • distributed and real-time

Architecture Overview

The SENSEI architecture with an Ambient wireless sensor network (WSN) is depicted in the figure below.

  • The localization DDI messages are received by the AmbientREP and the last position in this localization message can be fetched via the AmbientREP, the resource is a combination of the SmartPoint? above and the position sensor
  • The outlier detection DDI messages are received by the AmbientREP and the last outlier result can be fetched via the AmbientREP, the resource is a combination of the Outlier Detecting MicroRouter? above and the outlier sensor
  • The IMA DDI messages are received by the AmbientREP and the last IMA value can be fetched via the AmbientREP, the resource is a combination of the Inertia ProMove? node above and the Activity sensor

Current features

  • sensor get/post/subscribe
    • http://<ip:port>/rep/<node>/sensor/temp, sensor/humidity, sensor/pos for smartpoint nodes
    • http://<ip:port>/rep/<node>/sensor/activity for Inertia/UTwente ProMove? nodes
    • http://<ip:port>/rep/<node>/sensor/outlier for dedicated Outlier-detecting MicroRouters? by UTwente
  • parameter get for http://<ip:port>/rep/<node>/parameter/battery for smartpoint and infrastructure nodes
  • actuator get/post(setting) for http://<ip:port>/rep/<node>/actuator/led<number> for getting and setting the led status on the number'ed led of an infrastructure node (number=0, 1 or 2)
  • overview of nodes in http://<ip:port>/rep/
  • overview of methods per node in http://<ip:port>/rep/<node>
  • obtaining last DDI messages of certain type with http://<ip:port>/rep/<node>/ddi/<class_id>:<driver_id>:<function_id>
  • sending XML-formatted DDI message to a node by posting to http://<ip:port>/rep/<node>/ddi/

Open issues

  • Not so obvious to sent multiple sensor values, e.g. DDI IMA has 6 values of past minute activity with 10 second interval but only the last one is used by sensor/activity resource
  • Implementing dedicated configuration through AmbientREP is rather cumbersome, especially when it has multiple parameters, it would then involve posting XML documents.
    • Sending configuration as DDI message in XML format via AmbientREP is an approach that would work from applications
    • configuration via AmbientStudio? is the current modes operandi

Planned features

  • none


  • enable fetching and subscribing to temperature, humidity, and position information from Ambient Network via SENSEI RAI interface.
  • enable fetching and subscribing to activity measurements for UTwente activity detection invention via SENSEI RAI interface.
  • enable fetching and subscribing to outlier detection results for UTwente outlier detection invention via SENSEI RAI interface.

How to build AmbientREP

Obtaining other nodes and software

Ambient SmartPoint? 3000 T (Temperature) or T/H (temperature + humidity) can be used as position, temperature and humidity resources.

  • The radio chip firmware contains the DDI driver for auto-calibrated localization

Ambient MicroRouter? 3000

  • The MicroRouters? take care of establishing radio coverage via wireless mesh networking. The main purpose of these intelligent hubs is to deliver data reliably to where it is needed. These devices have extended radio range and are typically mounted within buildings or vehicles (e.g. trailers).

Outlier detecting Ambient MicroRouter? 3000

  • The radio chip software was updated by Ambient to forward SmartPoint? messages to the outlier detection DDI driver in the MSP.
  • The MSP430 software was updated by the University of Twente with an outlier detection DDI driver that receives the forwarded SmartPoint? messages from the radio chip and runs the Outlier detection algorithm on it.
    • The result of the algorithm is sent as a DDI message towards the gateway.
    • The Outlier detection algorithm can also be configured via the gateway using DDI messages

Ambient Gateway 3000

  • Each wireless network contains at least one Gateway, which ensures that data is communicated between the wireless network and corporate IT systems. To ease integration, Ambient introduced the Ambient Middleware, which “talks” XML over TCP/IP.

Ambient Middleware, developed by Ambient

  • AmbientStudio? contains the Ambient middleware and can be used to visualize sensor data, to display device locations on a map or to configure your SmartPoints? by setting rules for Dynamic Event Reporting™. This software product runs on a PC and can interface with ConnectBoxes? or directly with Gateways. Various plug-ins are available to interface with user-defined application or interfaces, as the SENSEI system.
  • the ConnectBox? contains the Ambient middleware and can be used for remote sensor monitoring and maintenance.

Activity Monitoring: Inertia ProMove? node

  • the radio chip firmware has been created by Ambient, and is similar to that of a SmartPoint? 3000 but adapted for working with the ProMove? node
  • the MSP430 firmware from Inertia Technology was updated by the University of Twente with:
    • software to calculate the IMA value every 10 seconds from the acceleration sensor data
    • a DDI driver that periodically sends the last 6 IMA values via the radio chip, and can also be requested to calculate and send the current IMA value

How to test it

A number of different nodes are required for different types of operation. For regular temperature or humidity sensing, the Ambient SmartPoint? 3000 T/H will do. For Outlier Detection the UT can provide the required MSP430 firmware for

Setting it up works as follows (AmbientREP has been tested on (Ubuntu) Linux + Java 1.6), first obtain and build it as described above. :

  • edit for setting ambient.rep.IP6 and ambient.rep.IP4 to proper DNS hostnames, and port for webserver with ambient.rep.port (default 8181)
  • edit for setting api.server.port to a free and reachable TCP port over IPv4 and/or IPv6 (organize NAT if necessary)
  • optionally start gogoc or gw6c client for ipv6 connectivity (e.g. with credentials for ambient), else use IPv4 address/hostname in weblinks below
  • run with: ant AmbientREP_APIserver
  • attach sensor networks with AmbientStudio? or ConnectBox? by specifying ConnectAPI_Client with hostname/IP address of machine running AmbientREP and TCP port api_server.port (as in second bullet)
  • point your web-browser to the configured hostname in first bullet, e.g., which will show active nodes in the ambient network
  • use SoapUI, FireFox? RESTClient plugin, or the SENSEI ManagementGUI to test posting subscriptions, the /rep/collector can be used to show the last notification

The testbed details of the AmbientREP can be found under testbedInfo?.

Last modified 13 years ago Last modified on Mar 2, 2011, 9:07:06 AM

Attachments (2)

Download all attachments as: .zip