Assumptions

1. The attributes Sensors.timestamp and Events.timestamp are used as timestamps of streams Sensors and Events, respectively.
   
   
2. Conceptually, each sensor streams all its readings once every instant of time. Any practical query evaluation technique clearly has to push some evaluation into the sensors. For example, in queries 1 and 2 below, the sampling operation can be pushed to the sensors so that only the sampled readings are streamed out of a sensor.

Queries

1. Data Collection Query: Report light readings from each sensor at the rate of one reading every 10 seconds.
   
   

   Select   Rstream (id, sample1(light) as light)
   From     Sensors [Range 10 Second
                     Slide 10 Second]
   Group By id
   

   
   Comments: sample1 is a user-defined relational-aggregate that returns one value uniformly at random from its input relation which in this case is the bag of light readings over each successive ten minute interval.
   
   
2. Light Monitoring Query: Sample data from each sensor at the rate of one reading every second and monitor the average light over the last 30 seconds.
   
   

   SampleStream: Select   Rstream (id, sample1(light) as light) 
                 From     Sensors [Range 1 Second
                                   Slide 1 Second]
                 Group By id
    
   AvgLight:     Select Avg(light)
                 From   SampleStream [Range 30 Second]
   

   
   Comments: For readability we have specified the subquery for producing the combined stream of one-second samples from all sensors (SampleStream) separately. The subquery for SampleStream uses the user-defined relational-aggregate sample1 from the Data Collection Query. The final result of the Light Monitoring Query is the derived relation AvgLight as specified above.
   
   
3. Event-Triggered Query: Whenever a sensor reports a bird-detection event, sample data from all sensors within 10 meters of the event at the rate of one reading every second for 10 seconds and report the average light and temperature.
   
   

   DelayedEventStream:
       Select Dstream (*)
       From   Events [Range 10 Second]
       Where  type = 'BIRD-DETECT'
   
   SampledSensorStream:
       Select   Rstream (id, location, sample1(light) as light, 
                         sample1(temperature) as temp)
       From     Sensors [Range 1 Second
                         Slide 1 Second]
       Group By id, location
   
   Result:
       Select   Rstream (E.location, Avg(light) as light, Avg(temp) as temp)
       From     DelayedEventStream [Now] E,
                SampledSensorStream [Range 10 Second] S
       Where    dist(E.location, S.location) < 10;
   

   
   Comments: This query has been specified using two subqueries. The first subquery produces DelayedEventStream which is the Event stream delayed by 10 seconds. Each event is delayed by 10 seconds to allow all the readings for the event to be made. The second subquery produces SampledSensorStream, the (sampled) stream consisting of one reading per sensor per second. As expressed, the sampled light and the sampled temperature may not correspond to a sensor reading at the same time instant; however, a different user-defined relational-aggregate can be used to solve this problem, if necessary. Also, the second subquery assumes that Sensor.location is functionally determined by Sensor.id. The outermost query uses the user-defined function dist that returns the distance in meters between two input locations.
   
   
4. Bird Counter Query: Keep a count of the number of times birds have entered a particular nest.
   
   

   Select  Count(*) 
   From    Events E
   Where   E.type = 'BIRD-ENTER' AND E.location = 'NEST-1012'
   

   
   
5. Delta Compression Query: Report the new light reading at a particular location every time it changes by 3.
   
   

   Select  Istream(delta_compr(light))
   From    Sensors
   Where   location = 'NEST-1012'
   

   
   Comments: delta_compr is a user-defined windowed-aggregate that outputs the new light reading every time it changes by 3.
   
   
6. Take Action Query: To keep birds and animals away from the sensors, apply a low electric pulse whenever an external body comes in contact with an active sensor on the field.
   
   

   Select generate_pulse(location)
   From   Events E
   Where  E.type = 'EXTERNAL-CONTACT'
   

   
   Comments: generate_pulse is a user-defined function that will make the sensor(s) at location generate a low electric pulse.