An overview of a project that explored the potential of lightweight, low-cost sensing in a commercial kitchen as a means to aid understanding of food safety compliance issues.

1. Using the Internet of Things for Food
Safety Compliance Checking
Milan Markovic, Peter Edwards, Martin Kollingbaum,
Computing Science, University of Aberdeen
Alan Rowe, Rowett Institute of Nutrition & Health,
University of Aberdeen

2. IoT & Commercial Kitchens
• Sensors
• Actuators
• Cloud-based
management
• Better food safety
• Waste reduction
• Energy & staffing costs
reduction
• Predictive maintenance
• Increased transparency
Motivation
Technologies

3. • The project explored the potential of lightweight, low-cost sensing
in a commercial kitchen as a means to aid understanding of food
safety compliance issues.
Project Background

4. Context
“It is the responsibility of people
producing and supplying food to
ensure it is safe and what it says it
is …”
HACCP (Hazard Analysis and Critical Control Point)
A system that helps food business operators look at
how they handle food and introduces procedures to
make sure the food produced is safe to eat.
Availability of low power, plastic
sensor tags , wireless meat
temperature probes.

5. CookSafe Manual

6. Potential Benefits of IoT Solution
• Government agencies such as FSS and FSA can better target
"problematic" businesses based on detailed sensor data logs leading
to a higher effectiveness of their limited resources
• Automated compliance monitoring of the food preparation processes
offers businesses additional reassurance that they are not causing any
harm to their customers

7. Objectives
• Deploy an IoT system capable of continuous temperature monitoring
of individual food items in a commercial kitchen
• Develop means to model the individual food items and associated
sensor data in the context of HACCP processes
• Using semantic web technologies, generate concise provenance
abstractions based on raw sensor data that can be used to prove
HACCP–based food safety compliance

8. Testbed
Rye & Soda
http://www.ryeandsoda. com/
Restaurant & bar located in Aberdeen city
centre.

9. Raw Sensor Data
• 1 burger kept in
chilled store and
then cooked
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
14:44:10 14:47:02 14:49:55 14:52:48 14:55:41 14:58:34
°C
Time
Wireless Tag
Meat Probe

10. FS-PROV - Food Safety Management Ontology

11. Provenance
• Provenance describes the use and production of entities by activities,
which may be influenced in various ways by agents (PROV-DM , W3C)
• Entity - a physical, digital, conceptual, or other kind of thing with
some fixed aspects
• Activity - something that occurs over a period of time and acts upon
or with entities
• Agent - something that bears some form of responsibility for an
activity taking place, for the existence of an entity, or for another
agent's activity

12. Provenance
Mix Bake
100 g
Sugar
1 Egg
100 g
Butter
Bob's
Cake
Batter
Cake
Bob
Mix Step
Cake
Batter
...Sugar
Constraint
90 - 100 g
...
Plan

13. FS-PROV
PROV-O
SC-PROV-O
P-PLAN FS-PROV
• Develop means to
model the
individual food
items and
associated sensor
data in the context
of HACCAP
processes

14. Generating Food Safety Compliance Records
HACCP Plan
Raw Sensor Data
Storage
Step
Preparation
Step
Cooking
Step
HACCP Chilled
Meat Constraint
<5 °C
HACCP Cooked
Meat Constraint
>75 °C
Chilled
Meat
Cooked
Meat
Meat
out of
Storage
Tag 1
Feature of
Interest
Burger 1
Surface
Temperature
4 °C Tag 1
Feature of
Interest
Burger 1
Surface
Temperature
7 °C
Meat
Probe
Feature of
Interest
Burger 1
Core
Temperature
76
°C
12 Jan 2018 15:40 12 Jan 2018 18:40 12 Jan 2018 18:45
Burger 1
(chilled)
True
12 Jan 2018 15:40 12 Jan 2018 18:40
Burger 1
(prepared)
12 Jan 2018 18:40
Burger 1
(cooked)
True
12 Jan 2018 18:45 12 Jan 2018 18:45
Preparation
Activity
Cooking
Activity
Storage
Activity

15. Combining Sensor Data
• 2 burgers cooked at
the same time
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
2:18:14 PM 2:21:07 PM 2:24:00 PM 2:26:53 PM 2:29:46 PM 2:32:38 PM 2:35:31 PM 2:38:24 PM 2:41:17 PM 2:44:10 PM
° C
Time
Meat Probe
Wireless Tag 1
Wireless Tag 2
Which
Burger
Cooked?

16. Conclusions
• It is possible to deploy IoT devices in a kitchen environment but many
challenges remain, including:
- Availability of network infrastructure
- Harsh physical environment
- Issues around sensor data quality/calibration
• Possible to identify HACCP compliance/non-compliance through
analysis of linked sensor data
• Provenance abstractions can represent compliance/non-compliance
within a scalable food safety management system
• Linking different sensor data streams to record a single food item
history is difficult

17. It is not Always a Straight Forward Story
“OMG – The sensor’s
melted!”
“The fridge has reached 9⚬C –
What on earth are they
doing!?”
“Burger core temperature of
150⚬C looks a bit
suspicious…”

18. New data -> new responsibilities....
• Most BIG DATA is stupid data and need to be put in a context
• IoT needs to be minimal effort as well as minimal cost
• Who is responsible for recording, storing, processing, and publishing
of the data?
• How is the data veracity assessed?
• Are the "things" really reporting the data that we need?

19. Thanks
Contact:
milan.markovic@abdn.ac.uk
p.edwards@abdn.ac.uk