Interpreting Exercise Data: Guidance and Coaching Modulearda.ai/.../10/Interpreting-Exercise-Data-Guidance-and-Coaching-Mod… · lower wattage indicating either a loss of fitness
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Interpreting Exercise Data: Guidance and Coaching Module *=Unique to Performance Lab
Performance Lab has the only context based AI exercise coaching system in the world.
Performance Lab’s software is able to:
• Detect the activity the user is engaged in,
• Compare the user’s specific activity data with other similar data,
• interpret the information and provide feedback. (Figure 1)
Figure 1. Map of Run Completed with Performance Lab’s AI Coaching Software. Each Dot of the Route Indicates a Coaching Comment Provided by the AI System with one Comment Highlighted to Show the kind of Feedback Supported.
Experiences: There are two possible guided exercise experiences:
• Guidance (metrics, gamified ‘winning the metric’, progress, in zone assistance)
• and coaching. (extra interpretative layer that suggests appropriate behaviour changes as a
coach would provide.)
Unique Functionality That Drives the Coaching Experience: Performance Lab’s system is the first to be able to truly coach a user for two reasons:
Being able to detect whether someone is walking, running, cycling or rowing makes for very coarse
identification of activity. Performance Lab’s patented algorithms1 allow much deeper insights to be
obtained due to its ability to contextually analyse a situation. The reason that detailed activity
identification has been very difficult until now is that exercise data is a tangle of interrelationships
between parameters2 (Figure 2):
Figure 2. Graph Showing Tangle of Relationships Between Heart Rate (Red), Speed (Dark Blue), Power (Light Blue), Stride Rate (Orange) and Terrain (Brown) During a Hilly Run.
Performance Lab’s technology can make sense of the data patterns by locating contextual signatures
in the data that identify an activity or sub activity (known as Training Types*)3. Within the mess of
multiple data parameters, when a match occurs, an activity is identified. (Figure 3) See: Detecting
Figure 5. Logged Data for Each Automatically Detected Activity Classification as: a) Individual Classifications Detected in a Timeline and b) a Summary of All the Data in the Workout for Each Classification.
The Classifications are also logged by location for future locational comparisons: (Figure 6)
Figure 6. Map of Simple Running Activities: a) Easy Running (green) and b) Hill Climbs (red).
The main benefit is that each Training type classification has its own unique and tightly constrained
set of criteria which means very detailed and accurate comparisons* can be made between current
data and historic data of the same classification4. (Figure 7)
Figure 12. Training Plan Showing that within the Workout, 2 Up Tempo Training Type Reps are Scheduled.
Once the runner is doing the workout, the Performance Lab system can detect all the different
Training Types automatically and keep track of them against the plan*. (See Comment’s in Figure 13)
Figure 13. Graphic Showing All the Commentary Dots with the gray vertical line just after a 2nd light blue block (depicting an Up Tempo Training Type Classification). The comment at this point shows that the system detected the Up Tempo rep, knew it was the 2nd occurrence in the workout, summarised the logged data for the segment (distance, duration, stride rate and average pace) which can then be compared to Training Plan Criteria.
After the Workout: After the workout (as shown in comments Figure 14), the coaching system knows what was
scheduled so it knows that the 54 mins of running was longer than planned. In this case, it also
knows that 2 Up Tempo reps of 30 seconds were scheduled. The ability to detect sub activities
(Training Types) within the workout means that the system could tell that 1 min 24s of Up Tempo
was over the prescribed amount but that the correct number of Up Tempo reps were completed*:
Figure 14. Graphic Showing Commentary at the End of the Runner’s Workout Indicating the Systems Knowledge of Both the Workout Schedule and Types of Training Types Performed Within the Workout.
Performance: Performance Lab’s proprietary Cardio Performance algorithms8 can also measure day to day Cardio
Performance*. (fitness)
During the Workout:
In each workout, the Performance algorithm measures a user in two ways; (Figure 15)
• the Cardio Performance Capacity (fitness) or
• the Cardio Performance Endurance (measures endurance and fatigue rate)
Figure 15. Cardio Performance Capacity and Cardio Performance Endurance measures.
Figure 18. Cardio Performance Measures Over Months of Exercise Exhibiting Effective Training, Overreaching (precursor to over training), and Overtraining.
Adapting the Plan: Finally, the exercise schedule can be adapted10 in real time* based on data that is measured and
interpreted during the workout. (as shown in Figure 19)
The training schedule can also be adapted post workout based on data collected.
Figure 19. Examples of Real Time In Workout Plan Adaptions.
Technique: The Coaching system can detect technique issues within a runner or cyclist’s activity. Figure 20
shows the coaching feedback around running at the correct stride rate: