Simple System
Easy to Deploy
Reduces Labor
Improves Performance
A flexible Artificial Intelligence (AI) and sensor system that can be deployed at any industrial process step to see and predict objects and events, recommend actions and report on important metrics. Its AI can be trained to make human-like decisions consistently and quickly.
An Agent for your industry, an Agent for your specific application
Food Production: Monitor appearance, size, and color of baked goods in real-time
Wood Products: Identify bottlenecks prior to equipment failure
Electronics: Detect misaligned components, soldering issues, and missing parts
Pharma: Ensure that production steps are recorded accurately for compliance
Plastics: Identify defective products prior to shipment
Fragrance: Adjust production temperatures and mixing times
How RIOS Agents Deliver Value
RIOS Agents: What They Do
RIOS Agents use exceptional visual intelligence capabilities to optimize production processes and reduce situations that can negatively impact production. Here's what they do:
- Detect People and Objects
RIOS Agents leverage Artificial Intelligence to train models that can detect people and nonuniform objects such as the top sheet of veneer panels. These models can detect simple and complex anomalies. - Detect Events
RIOS Agents can be trained to detect events that can stop manufacturing processes - reducing downtime and increasing reliability.
- Make Predictions
RIOS Agents can predict issues before they happen - prompting staff to intervene prior to a machine failure. They can prompt staff intervention and stop or slow conveyers. - Recommend or Execute Actions
RIOS Agents can recommend or execute actions such as starting, stopping or adjusting the speed of machinery, scheduling maintenance, or a variety of other productive actions.
Applications for RIOS Agents
Examples of process steps by industry where RIOS Agents add value
Location | Value |
Log Singulation | Reduced costs due to doubles |
Lathe Infeed | Eliminate or reduce lathe tendre labor |
Clipper/Chipper Infeed | Eliminate downtime at machine centers clearing jams |
Dryer Infeed | Eliminate Headcount at Dryer Infeed |
Dryer Outfeed | Clear doubles, wads off dryer infeed prior to entry into grade line. Eliminate headcount at 90 degree turns |
Layup line | Increase line throughput, quality of output and reduce waste from downstream |
Core/99” Saw Infeed | Prevent Core saws from being loaded incorrectly |
Sander | Identify line management of sander related defects |
Package Quality | Reduced quality labor, increased value |
Unauthorized entry | More secure and safe environment, reduced incidents |
Stock monitoring | Better stock control |
Location | Value |
Debarker | Eliminate downtime/outages tied to debarker damage and log extraction |
Planar feeder | More efficient planar operation with higher quality output |
Unscrambler, Lug loader,Trimmer infeed | Better lug fill, eliminate labor on unscrambler outfeed better trim saw performance |
Better quality outfeed, prevent jams in conveyance to round table/grader | |
Round table | Reduced labor cost, consistent sortation and rejection with quality details and standards |
Stick separator | More efficient and effective separation |
Sorter | Stopping line in time to address torpedo board |
Kiln | Eliminates kiln damage, replaces wire based board detection |
Stock monitoring | Better stock control |
Unauthorized entry | Proactive defense for unauthorized entry |
Location | Value |
Safety | Fewer incidents |
Forming line | More uptime, lower maintenance costs due to fewer and less severe incidents |
Hot press | More uptime, lower maintenance costs due to fewer and less severe incidents |
Plant outfeed | Increased quality and reliability of output due to enhanced feedback Lower costs due to quality issues due to same |
Debarker infeed | Reduced oversight labor Fewer incidents, more uptime, less maintenance Higher quality and more consistent input |
Log grappling from pond | Less downtime, higher quality due to better anomalous log handling |
Location | Value |
Moulder | Better preventative saw maintenance, less waste
Less skilled labor needed to manage outfeed |
Rip saw | Increased quality and better maintenance
Reduced labor required per board processed |
Unauthorized entry | More secure and safe environment, reduced incidents |
Stock monitoring | Better stock control |
| Process Step | AI Vision Target(s) | Potential Automation | Value |
| Pack to box | Label quality, expiration date, lot code | Notification, stop, slow, robotic fix | Prevent products with damaged labels or missing expiration dates from being shipped to customer |
| Pack to thermoformer | Food product fit into package | Notification, stop, slow, robotic fix | Prevent air leaks that can lead to spoilage of food product |
| Process Step | AI Vision Target(s) | Potential Automation | Value |
| Queing for palletizing | Box integrity | Notification, stop, slow, robotic fix | Prevent damaged goods from being shipped to customer |
| Crate cleaning | Crate integrity inspection | Notification, stop, slow, robotic fix | Prevent damaged crates from being used which could damage product |
Applications for Intelligent Agents
Examples of process steps by industry where Intelligent Agents add value
| Process Step | AI Vision Target(s) | Potential Automation | Value |
| Trim saw outfeed | Skew, double lug | Notification, stop, slow, robotic fix | Better quality outfeed, prevent jams in conveyance to round table/grader… |
| Stick separator | Stick full length without broken ends | Notification, robotic fix | More efficient and effective separation |
| Location | AI Vision Target(s) | Potential Automation | Value |
| Log Singulation | Doubles | Notify, Stop, Fix | Reduced costs due to doubles |
| Lathe Infeed | Lathe tender | Notify, Stop, Slow, Fix | Eliminate or reduce lathe tendre labor |
| Clipper/Chipper Infeed | Wad management | Slow, Fix | Eliminate downtime at machine centers clearing jams |
| Dryer Infeed | Loader misfeed | Identify, Fix | Eliminate Headcount at Dryer Infeed |
| Dryer Outfeed | Multi sheet stack to grader | Notify, Stop, Fix | Clear doubles, wads off dryer infeed prior to entry into grade line. Eliminate headcount at 90 degree turns. |
| Core/99” Saw Infeed | Load correction | Stop | Prevent Core saws from being loaded incorrectly |
| Sander | Skips, mars | Notify | Identify line management of sander related defects |
| Package Quality | Quality issue detection and action | Notify | Reduced quality labor, increased value |
| Unauthorized entry | Access by people | Notify, Stop | More secure and safe environment, reduced incidents |
| Stock monitoring | stock levels | Notify, lockdown | Better stock control |
| Location | AI Vision Target(s) | Potential Automation | Value |
| Debarker | Double logs | AI supervisor to notify of double log entry | Eliminate downtime/outages tied to debarker damage and log extraction |
| Planar feeder | Crossups, double lugs, boards not even | Notification, stop, robotic fix | More efficient planar operation with higher quality output |
| Unscrambler, Trimmer infeed | Doubles, crossups, boards not even | Notification on double board. Robotic response to reject double. | Better lug fill, eliminate labor on unscrambler outfeed better trim saw performance. |
| Trim saw outfeed | Skew, double lug | Notification, stop, slow, robotic fix | Better quality outfeed, prevent jams in conveyance to round table/grader… |
| Stick separator | Stick full length without broken ends | Notification, robotic fix | More efficient and effective separation |
| Sorter | Hangups | Torpedo board notification | Stopping line in time to address torpedo board. |
| Kiln | Hanging Boards on Package | AI supervisor to stop package entry into CDK kils for packages with handing boards | Eliminates kiln damage, replaces wire based board detection. |
| Stock monitoring | stock levels | Notify, lockdown | Better stock control |
| Unauthorized entry | People entering LOTO space | AI supervisor to issue all stop to PLC. Notification to supervisor | Proactive defence for unauthorized entry. |
| Location | AI Vision Target(s) | Potential Automation | Value | Log grappling from pond | Anomalous logs Log grappling quality/incidents | Notify of anomalous logs Grappling guidance Grappling automation | Less downtime, higher quality due to better anomalous log handling |
| Debarker infeed | Anomalous logs Empties/doubles | Notify problems for human intervention Slow/stop line before incident Robotic intervention | Reduced oversight labor Fewer incidents, more uptime, less maintenance Higher quality and more consistent input |
| Forming line | Contamination | Notify operator of potential incident Stop before incident | More uptime, lower maintenance costs due to fewer and less severe incidents |
| Hot press | Jam or default movement of carger | Notify operator Stop before incident | More uptime, lower maintenance costs due to fewer and less severe incidents |
| Plant outfeed | Product quality | Notify operators of quality issues Signal operators/machines earlier in process of issues in real time for action Robotic handling of quality issues | Increased quality and reliability of output due to enhanced feedback Lower costs due to quality issues due to same |
| Process Step | AI Vision Target(s) | Potential Automation | Value |
| Queing for palletizing | Box integrity | Notification, stop, slow, robotic fix | Prevent damaged goods from being shipped to customer |
| Crate cleaning | Crate integrity inspection | Notification, stop, slow, robotic fix | Prevent damaged crates from being used which could damage product. |
| Process Step | AI Vision Target(s) | Potential Automation | Value |
| Pack to box | Label quality, expiration date, lot code | Notification, stop, slow, robotic fix | Prevent products with damaged labels or missing expiration dates from being shipped to customer |
| Pack to thermoformer | Food product fit into package | Notification, stop, slow, robotic fix | Prevent air leaks that can lead to spoilage of food product |
RIOS Agents for Wood Products
DEFECT DETECTION at layup infeed
RIOS Agents can be used to identify characteristics in wood products, such as knots and splits. In this example, the agent identifies holes, knots, and splits and provides a probability for each event. Plywood mills leverage this information to reject or recycle veneer and report on quality.
BACKLOG DETECTION
RIOS Agents are used for monitoring backlog. The rectangular zone indicates the region under observation. This zone actively monitors the current number of studs available in the backlog/buffer area.
provide new metrics
RIOS Agents provide analytics and reporting back to Mill operations staff to track key metrics for their products. Leveraging computer vision, the agent in this example is counting veneer panels and the number of knots, holes, splits.
Count bottles of various sizes and shapes that are in containers as they move down your conveyors.
A simple way to use Mission Control is to identify what a bottle cap looks like by drawing bounding boxes around the caps of each bottle in a sample image to help the algorithm learn what they look like. The system then creates a machine-learning model that can identify bottles in containers as they move down your conveyer line.
You can leverage Mission Control’s Auto Detection feature to find all of the caps and train your machine learning model automatically. You can test your models, side by side, to see how they perform.
You can build and compare models in minutes. You can compare them side by side to view their performance.
Mission Control is a no-code platform that enables manufacturers to view videos, identify objects of interest, and automatically build and update learning models that provide deep introspection and control of their processes. You don’t need to learn AI or Machine Language.
Using video feeds, operators can open the application and create models instantly. You can identify the objects of interest, detect defects on objects, and identify how objects are handled – all with a point-and-click process that requires no coding or complex training.
RIOS Agents lower the barrier of introducing reliable and effective machine learning. Manufacturers can easily build complex processes using simple, easy-to-use web interfaces. These processes can do everything from control complex robotics movements to identifying objects.